Author: David Clark

Economy-Watching: U.S. Market Probability Tracker Updated with June’s Employment Data

[from the Federal Reserve Bank of Atlanta]

The following information is now available on the Federal Reserve Bank of Atlanta’s website.

Market Probability Tracker Updated with New Employment Data
On Friday, the U.S. Bureau of Labor Statistics released June’s employment report. Find out how this affected the market’s assessment of future rate moves at the Market Probability Tracker.

Global Supply Chain Pressure Index: July 2022 Update

[from the Federal Reserve Bank of New York, Applied Macroeconomics and Econometrics Center]

A new reading of the Global Supply Chain Pressure Index has been posted.

The GSCPI compiles more than two dozen metrics across seven economies—data on global transportation costs and regional manufacturing conditions—to track shifts in supply chain pressures from 1997 to the present.

The GSCPI is updated regularly at 10:00am ET on the fourth business day of each month.

Estimates for June 2022
  • Global supply chain pressures declined in June, continuing the decrease we observed for May.
  • The June decline was mostly due to a large decrease in Chinese supply delivery times.
  • The moves in the GSCPI over the past three months suggest that although global supply chain pressures have been decreasing, they remain at historically high levels.

The GSCPI is a product of the Federal Reserve Bank of New York’s Applied Macroeconomics and Econometrics Center.

View the Index.

Science-Watching: Forecasting New Diseases in Low-Data Settings Using Transfer Learning

[from London Mathematical Laboratory]

by Kirstin Roster, Colm Connaughton & Francisco A. Rodrigues

Abstract

Recent infectious disease outbreaks, such as the COVID-19 pandemic and the Zika epidemic in Brazil, have demonstrated both the importance and difficulty of accurately forecasting novel infectious diseases. When new diseases first emerge, we have little knowledge of the transmission process, the level and duration of immunity to reinfection, or other parameters required to build realistic epidemiological models. Time series forecasts and machine learning, while less reliant on assumptions about the disease, require large amounts of data that are also not available in early stages of an outbreak. In this study, we examine how knowledge of related diseases can help make predictions of new diseases in data-scarce environments using transfer learning. We implement both an empirical and a synthetic approach. Using data from Brazil, we compare how well different machine learning models transfer knowledge between two different dataset pairs: case counts of (i) dengue and Zika, and (ii) influenza and COVID-19. In the synthetic analysis, we generate data with an SIR model using different transmission and recovery rates, and then compare the effectiveness of different transfer learning methods. We find that transfer learning offers the potential to improve predictions, even beyond a model based on data from the target disease, though the appropriate source disease must be chosen carefully. While imperfect, these models offer an additional input for decision makers for pandemic response.

Introduction

Epidemic models can be divided into two broad categories: data-driven models aim to fit an epidemic curve to past data in order to make predictions about the future; mechanistic models simulate scenarios based on different underlying assumptions, such as varying contact rates or vaccine effectiveness. Both model types aid in the public health response: forecasts serve as an early warning system of an outbreak in the near future, while mechanistic models help us better understand the causes of spread and potential remedial interventions to prevent further infections. Many different data-driven and mechanistic models were proposed during the early stages of the COVID-19 pandemic and informed decision-making with varying levels of success. This range of predictive performance underscores both the difficulty and importance of epidemic forecasting, especially early in an outbreak. Yet the COVID-19 pandemic also led to unprecedented levels of data-sharing and collaboration across disciplines, so that several novel approaches to epidemic forecasting continue to be explored, including models that incorporate machine learning and real-time big data data streams. In addition to the COVID-19 pandemic, recent infectious disease outbreaks include Zika virus in Brazil in 2015, Ebola virus in West Africa in 2014–16, Middle East respiratory syndrome (MERS) in 2012, and coronavirus associated with severe acute respiratory syndrome (SARS-CoV) in 2003. This trajectory suggests that further improvements to epidemic forecasting will be important for global public health. Exploring the value of new methodologies can help broaden the modeler’s toolkit to prepare for the next outbreak. In this study, we consider the role of transfer learning for pandemic response.

Transfer learning refers to a collection of techniques that apply knowledge from one prediction problem to solve another, often using machine learning and with many recent applications in domains such as computer vision and natural language processing. Transfer learning leverages a model trained to execute a particular task in a particular domain, in order to perform a different task or extrapolate to a different domain. This allows the model to learn the new task with less data than would normally be required, and is therefore well-suited to data-scarce prediction problems. The underlying idea is that skills developed in one task, for example the features that are relevant to recognize human faces in images, may be useful in other situations, such as classification of emotions from facial expressions. Similarly, there may be shared features in the patterns of observed cases among similar diseases.

The value of transfer learning for the study of infectious diseases is relatively under-explored. The majority of existing studies on diseases remain in the domain of computer vision and leverage pre-trained neural networks to make diagnoses from medical images, such as retinal diseases, dental diseases, or COVID-19. Coelho and colleagues (2020) explore the potential of transfer learning for disease forecasts. They train a Long Short-Term Memory (LSTM) neural network on dengue fever time series and make forecasts directly for two other mosquito-borne diseases, Zika and Chikungunya, in two Brazilian cities. Even without any data on the two target diseases, their model achieves high prediction accuracy four weeks ahead. Gautam (2021) uses COVID-19 data from Italy and the USA to build an LSTM transfer model that predicts COVID-19 cases in countries that experienced a later pandemic onset.

These studies provide empirical evidence that transfer learning may be a valuable tool for epidemic forecasting in low-data situations, though research is still limited. In this study, we aim to contribute to this empirical literature not only by comparing different types of knowledge transfer and forecasting algorithms, but also by considering two different pairs of endemic and novel diseases observed in Brazilian cities, specifically (i) dengue and Zika, and (ii) influenza and COVID-19. With an additional analysis on simulated time series, we hope to provide theoretical guidance on the selection of appropriate disease pairs, by better understanding how different characteristics of the source and target diseases affect the viability of transfer learning.

Zika and COVID-19 are two recent examples of novel emerging diseases. Brazil experienced a Zika epidemic in 2015–16 and the WHO declared a public health emergency of global concern in February 2016. Zika is caused by an arbovirus spread primarily by mosquitoes, though other transmission methods, including congenital and sexual have also been observed. Zika belongs to the family of viral hemorrhagic fevers and symptoms of infection share some commonalities with other mosquito-borne arboviruses, such as yellow fever, dengue fever, or chikungunya. Illness tends to be asymptomatic or mild but can lead to complications, including microcephaly and other brain defects in the case of congenital transmission.

Given the similarity of the pathogen and primary transmission route, dengue fever is an appropriate choice of source disease for Zika forecasting. Not only does the shared mosquito vector result in similar seasonal patterns of annual outbreaks, but consistent, geographically and temporally granular data on dengue cases is available publicly via the open data initiative of the Brazilian government.

COVID-19 is an acute respiratory infection caused by the novel coronavirus SARS-CoV-2, which was first detected in Wuhan, China, in 2019. It is transmitted directly between humans via airborne respiratory droplets and particles. Symptoms range from mild to severe and may affect the respiratory tract and central nervous system. Several variants of the virus have emerged, which differ in their severity, transmissibility, and level of immune evasion.

Influenza is also a contagious respiratory disease that is spread primarily via respiratory droplets. Infection with the influenza virus also follows patterns of human contact and seasonality. There are two types of influenza (A and B) and new strains of each type emerge regularly. Given the similarity in transmission routes and to a lesser extent in clinical manifestations, influenza is chosen as the source disease for knowledge transfer to model COVID-19.

For each of these disease pairs, we collect time series data from Brazilian cities. Data on the target disease from half the cities is retained for testing. To ensure comparability, the test set is the same for all models. Using this empirical data, as well as the simulated time series, we implement the following transfer models to make predictions.

  • Random forest: First, we implement a random forest model which was recently found to capture well the time series characteristics of dengue in Brazil. We use this model to make predictions for Zika without re-training. We also train a random forest model on influenza data to make predictions for COVID-19. This is a direct transfer method, where models are trained only on data from the source disease.
  • Random forest with TrAdaBoost: We then incorporate data from the target disease (i.e., Zika and COVID-19) using the TrAdaBoost algorithm together with the random forest model. This is an instance-based transfer learning method, which selects relevant examples from the source disease to improve predictions on the target disease.
  • Neural network: The second machine learning algorithm we deploy is a feed-forward neural network, which is first trained on data of the endemic disease (dengue/influenza) and applied directly to forecast the new disease.
  • Neural network with re-training and fine-tuning: We then retrain only the last layer of the neural network using data from the new disease and make predictions on the test set. Finally, we fine-tune all the layers’ parameters using a small learning rate and low number of epochs. These models are examples of parameter-based transfer methods, since they leverage the weights generated by the source disease model to accelerate and improve learning in the target disease model.
  • Aspirational baseline: We compare these transfer methods to a model trained only on the target disease (Zika/COVID-19) without any data on the source disease. Specifically, we use half the cities in the target dataset for training and the other half for testing. This gives a benchmark of the performance in a large-data scenario, which would occur after a longer period of disease surveillance.

The remainder of this paper is organized as follows. The models are described in more technical detail in Section 2. Section 3 shows the results of the synthetic and empirical predictions. Finally, Section 4 discusses practical implications of the analyses.

Access the full paper [via institutional access or paid download].

FDIC News: Agencies Issue Host State Loan-to-Deposit Ratios

[from Federal Deposit Insurance Corporation, released June 28]

Federal bank regulatory agencies today issued the host state loan-to-deposit ratios that are used to evaluate compliance with section 109 of the Riegle-Neal Interstate Banking and Branching Efficiency Act of 1994. These ratios replace those from June 2021.

By law, a bank is generally prohibited from establishing or acquiring branches outside of its home state primarily for the purpose of acquiring additional deposits. This prohibition seeks to ensure that interstate bank branches will not take deposits from a community without the bank also reasonably helping to meet the credit needs of that community.

Section 109 Host State Loan-to-Deposit Ratios [Archived PDF].

Nature Portfolio Collection: Nature Mental Health

[from Nature]

To mark the upcoming launch of Nature Mental Health, and to showcase the potential breadth and scope of the journal, the editors present a collection of recent and representative mental health-themed articles from across the Nature portfolio.

In this Nature portfolio collection:

Read the full collection [some articles require a subscription].

Russia-Watching: Economic Dysfunctionalities

[from the Russian Analytical Digest]

This issue deals with dysfunctionalities in the Russian economy. The first three contributions look at the direct impact of sanctions. Ilya Matveev provides an overview, while Andrei Yakovlev compares the government’s anti-sanctions measures to its reaction to the economic impact of the COVID-19 pandemic. Janis Kluge offers a more detailed picture of the short- and long-term effects of the unfolding sanction regime. Michael Rochlitz then goes on to explain the lack of strategic planning in the country’s economic policy. Finally, Olga Masyutina and Ekaterina Paustyan provide a case study of inefficient governance mechanisms looking at waste management.

Read the full issue [archived PDF].

Analyses

Sanctions against Russia: No Blitzkrieg, but a Devastating Effect Nonetheless

by Ilya Matveev

In response to the Russian invasion of Ukraine, over 40 countries have introduced sanctions against Russia. The new restrictions concern finance, trade, logistics, and personal sanctions against businessmen and officials. In addition, more than 1,000 companies have ceased or limited their activities in Russia. In this article, Ilya Matveev argues that the sanctions, despite their unprecedented scale, have not led to the collapse of the Russian economy, yet their effect is dramatic, multi-faceted, and will increase over time.

Read the full issue [archived PDF].

Fighting the Pandemic and Fighting Sanctions: Can the Russian Economy Now Benefit from Its Experience with Anti-Crisis Measures?

by Andrei Yakovlev

Faced with tough international sanctions in reaction to its war against Ukraine, the Russian government has resorted to measures developed during the COVID-19 pandemic in order to stabilize the economy. This short analysis discusses the rationale behind this approach and demonstrates its limits.

Read the full issue [archived PDF].

Russia’s Economy under Sanctions: Early Impact and Long-Term Outlook

by Janis Kluge

Four months after a coalition of Western states imposed unprecedented sanctions on Russia, the Russian economy seems to be holding up better than expected. The Central Bank has managed to stabilize the country’s financial system and Russian officials are trying to project optimism about the future. However, this optimism is likely to be short-lived. The sanctions’ effects are only just beginning to unfold: supply-chain problems are intensifying and demand is falling quickly. In the longer run, Russia’s economy will become more primitive as it partially decouples from international trade. To avoid social tensions, the government will intervene to support Russian businesses, leading to more protectionism and a larger state footprint in the economy.

Read the full issue [archived PDF].

Why Russia Is Lacking an Economic Strategy for the Future

by Michael Rochlitz

Even before the economic crisis caused by Russia’s full-scale attack against Ukraine and the ensuing sanctions, the Russian economy was plagued by a number of growing problems. As a result, Russia’s economy has hardly grown for almost a decade, with an average annual growth rate of just 0.5% between 2013 and 2021. However, the Russian government does not have a strategy for addressing the fundamental economic challenges that are looming just over the horizon. There also seem to be no public debates about these challenges, whether in the policy circles around the government or among the wider public.

Read the full issue [archived PDF].

The Political Economy of Waste Management in Russia

by Olga Masyutina and Ekaterina Paustyan

The problem of household waste is one of the numerous environmental challenges facing Russia today. The 2019 nation-wide waste management reform was designed to tackle this problem by promoting recycling. However, the reform is stalling, due in large part to the nature of state-business relations in Russia. The lack of transparency in the public procurement process and the importance of personal connections between businesses and the federal and regional authorities undermine the implementation of the reform and produce suboptimal outcomes in the fight against waste.

Read the full issue [archived PDF].

Coronavirus Update: Fall Boosters Could Have Bits of Omicron

[from ScienceNews Coronavirus Update, by Erin Garcia de Jesús]

For all the coronavirus variants that have thrown pandemic curve balls—including alpha, beta, gamma, delta—COVID-19 vaccines have stayed the same. That could change this fall.

Yesterday, an advisory committee to the U.S. Food and Drug Administration met to discuss whether vaccine developers should update their jabs to include a portion of the omicron variant—the version of the coronavirus that currently dominates the globe. The verdict: The omicron variant is different enough that it’s time to change the vaccines. Exactly how is up in the air; the FDA still has to weigh in and decide what versions of the coronavirus will be in the shot.

“This doesn’t mean that we are saying that there will be boosters recommended for everyone in the fall,” Amanda Cohn, chief medical officer for vaccine policy at the U.S. Centers for Disease Control and Prevention said at the June 28 advisory meeting. “But my belief is that this gives us the right vaccine for preparation for boosters in the fall.”

The decision to update COVID-19 vaccines didn’t come out of nowhere. In the two-plus years that the coronavirus has been spreading around the world, it has had a few “updates” of its own—mutating some of its proteins that allow the virus to more effectively infect our cells or hide from our immune systems.

Vaccine developers had previously crafted vaccines to tackle the beta variant that was first identified in South Africa in late 2020. Those were scrapped after studies showed that current vaccines remained effective.

The current vaccines gave our immune systems the tools to recognize variants such as beta and alpha, which each had a handful of changes from the original SARS-CoV-2 virus that sparked the pandemic. But the omicron variant is a slipperier foe. Lots more viral mutations combined with our own waning immunity mean that omicron can gain a foothold in the body. And vaccine protection isn’t as good as it once was at fending off COVID-19 symptoms.

The shots still largely protect people from developing severe symptoms, but there has been an uptick in hospitalizations and deaths among older age groups, Heather Scobie, deputy team lead of the CDC’s Surveillance and Analytics Epidemiology Task Force said at the meeting. And while it’s impossible to predict the future, we could be in for a tough fall and winter, epidemiologist Justin Lessler of the University of North Carolina at Chapel Hill said at the meeting. From March 2022 to March 2023, simulations project that deaths from COVID-19 in the United States might number in the tens to hundreds of thousands.

A switch to omicron-containing jabs may give people an extra layer of protection for the upcoming winter. Pfizer–BioNTech presented data at the meeting showing that updated versions of its mRNA shot gave clinical trial participants a boost of antibodies that recognize omicron. One version included omicron alone, while the other is a twofer, or bivalent, jab that mixes the original formulation with omicron. Moderna’s bivalent shot boosted antibodies too. Novavax, which developed a protein-based vaccine that the FDA is still mulling whether to authorize for emergency use, doesn’t have an omicron-based vaccine yet, though the company said its original shot gives people broad protection, generating antibodies that probably will recognize omicron.

Pfizer and Moderna both updated their vaccines using a version of omicron called BA.1, which was the dominant variant in the United States in December and January. But BA.1 has siblings and has already been outcompeted by some of them.

Since omicron first appeared late last year, “we’ve seen a relatively troubling, rapid evolution of SARS-CoV-2,” Peter Marks, director of the FDA’s Center for Biologics Evaluation and Research in Silver Spring, Maryland, said at the advisory meeting.

Now, omicron subvariants BA.2, BA.2.12.1, BA.4 and BA.5 are the dominant versions in the United States and other countries. The CDC estimates that roughly half of new U.S. infections the week ending June 25 were caused by either BA.4 or BA.5. By the time the fall rolls around, yet another new version of omicron—or a different variant entirely—may join their ranks. The big question is which of these subvariants to include in the vaccines to give people the best protection possible.

BA.1, the version already in the updated vaccines, may be the right choice, virologist Kanta Subbarao said at the FDA meeting. An advisory committee to the World Health Organization, which Subbarao chairs, recommended on June 17 that vaccines may need to be tweaked to include omicron, likely BA.1. “We’re not trying to match [what variants] may circulate,” Subbarao said. Instead, the goal is to make sure that the immune system is as prepared as possible to recognize a wide variety of variants, not just specific ones. The hope is that the broader the immune response, the better our bodies will be at fighting the virus off even as it evolves.

The variant that is farthest removed from the original virus is probably the best candidate to accomplish that goal, said Subbarao, who is director of the WHO’s Collaborating Center for Reference and Research on Influenza at the Doherty Institute in Melbourne, Australia. Computational analyses of how antibodies recognize different versions of the coronavirus suggest that BA.1 is probably the original coronavirus variant’s most distant sibling, she said.

Some members of the FDA advisory committee disagreed with choosing BA.1, instead saying that they’d prefer vaccines that include a portion of BA.4 or BA.5. With BA.1 largely gone, it may be better to follow the proverbial hockey puck where it’s going rather than where it’s been, said Bruce Gellin, chief of Global Public Health Strategy with the Rockefeller Foundation in Washington, D.C. Plus, BA.4 and BA.5 are also vastly different from the original variant. Both BA.4 and BA.5 have identical spike proteins, which the virus uses to break into cells and the vaccines use to teach our bodies to recognize an infection. So when it comes to making vaccines, the two are somewhat interchangeable.

There are some real-world data suggesting that current vaccines offer the least amount of protection from BA.4 and BA.5 compared with other omicron subvariants, Marks said. Pfizer also presented data showing results from a test in mice of a bivalent jab with the original coronavirus strain plus BA.4/BA.5. The shot sparked a broad immune response that boosted antibodies against four omicron subvariants. It’s unclear what that means for people.

Not everyone on the FDA advisory committee agreed that an update now is necessary—two members voted against it. Pediatrician Henry Bernstein of Zucker School of Medicine at Hofstra/Northwell in Uniondale, N.Y., noted that the current vaccines are still effective against severe disease and that there aren’t enough data to show that any changes would boost vaccine effectiveness. Pediatric infectious disease specialist Paul Offit of Children’s Hospital of Philadelphia said that he agrees that vaccines should help people broaden their immune responses, but he’s not yet convinced omicron is the right variant for it.

Plenty of other open questions remain too. The FDA could authorize either a vaccine that contains omicron alone or a bivalent shot, although some data hinted that a bivalent dose might spark immunity that could be more durable. Pfizer and Moderna tested their updated shots in adults. It’s unclear what the results mean for kids. Also unknown is whether people who have never been vaccinated against COVID-19 could eventually start with such an omicron-based vaccine instead of the original two doses.

Maybe researchers will get some answers before boosters start in the fall. But health agencies need to make decisions now so vaccine developers have a chance to make the shots in the first place. Unfortunately, we’re always lagging behind the virus, said pediatrician Hayley Gans of Stanford University. “We can’t always wait for the data to catch up.”

OFR Working Paper Finds Cash Biases Measurement of the Stock Return Correlations

[from the U.S. Office of Financial Research]

Today, the U.S. Office of Financial Research published a working paper, “Cash-Hedged Stock Returns” [archived PDF], and an accompanying blog (below), regarding firms’ cash holdings and the implications for asset prices and financial stability.

Cash holdings are important for financial stability because of their value in crises.  Corporate cash piles vary across companies and over time. Firms’ cash holdings typically earn low returns, and their cash returns are correlated across firms.  Thus, the asset pricing results are important for investors managing a portfolio’s risk and policymakers concerned about sources of vulnerability.

The working paper [archived PDF] shows how investors can hedge cash on firms’ balance sheets when making portfolio choices.  Cash generates variation in beta estimates, and the working paper decomposes stock betas into components that depend on the firm’s cash holding, return on cash, and cash-hedged return. Common asset pricing premia have large implicit cash positions, and portfolios of cash-hedged premia often have higher Sharpe ratios, used by investors to understand a return on investment, because of the correlation between firms’ cash returns. The paper shows the value of a dollar increased in 2020, and firms hold cash because they are riskier.

Read the working paper [archived PDF].

OFR Finds Large Cash Holdings Can Lead to Mismeasuring Risk

[from the OFR blog, by Sharon Ross]

Cash is necessary for companies’ operations. Firms use cash to make payments, finance investments, and manage risk. But holding cash comes at a cost: its low pecuniary return. Published today by the OFR, the working paper, “Cash-Hedged Stock Returns” [archived PDF], shows that the cash returns of publicly traded, non-financial firms are correlated. Since cash returns are a part of equity returns, investors that are using equity return correlations to measure risk can mismeasure risk.

We show the importance of cash for systemic risk by documenting the value of cash in crises, showing that firms hold cash in part due to risk management and studying how cash biases the measurement of the interconnectedness of stock returns. The consequences of cash are important for policymakers monitoring aggregate risks, and sources of market vulnerability and for investors making portfolio choices.

Cash holdings are important for financial stability because of their value in crises. Several papers document a “dash for cash” during the initial panicked stages of the coronavirus 2019 (COVID-19) pandemic when firms rushed to hold cash in their coffers. The dash for cash was driven by firms drawing down on lines of credit from banks, which in turn affected bank lending. The dash for cash highlighted the critical role of firms’ cash holdings and returns in understanding risk in the financial system.

We show the value of a dollar increased in 2020. Moreover, our results show that firms may hold cash because they are riskier, as opposed to firms with high cash shares being less risky due to their cash holdings. Our results are consistent with a precautionary savings motive for holding cash. In other words, firms hold cash for risk management, in part to weather bad times.

Cash is a growing share of public firms’ assets. The value-weighted U.S. stock market held 22% of its assets in cash in December 2020 compared to 8% in the 1980s. An investor buying the market in 2020 ends up with an implicit cash position three times larger than in 1980. Individual firms vary in how much cash they hold. As cash holdings increase, it is important to understand how cash holdings affect returns, which in turn impacts who chooses to invest in the firms.

Cash returns are correlated across firms, and cash biases measurement of the interconnectedness of stock returns, making it a risk for financial stability. As a result, the asset pricing results are important both for investors managing portfolio risk and for policymakers concerned about interconnected returns.

We argue that the value of corporate cash is distinct, and we can separate the value of cash and the value of the firm’s primary business. We show how investors can explicitly account for the effect of corporate cash holdings when forming a portfolio. When an investor owns stock in a company with substantial cash, the investor has an implicit cash position managed by the company—something the investor might not intend. We argue that investors should account for the effect of corporate cash holdings in the portfolio decision to measure a portfolio’s risk. Firms’ cash management is not consistent across firms, and investors may want to manage their cash positions themselves. Policymakers should be aware of investors’ choices in cash because of investors’ portfolio risk and the implications for aggregate risk.

We separate a company’s stock return into its cash and non-cash components, and we show that using the non-cash return gives a more informative correlation structure across stocks. In other words, if investors take out the correlated cash returns, the remaining return is less correlated, yielding portfolios that provide better diversification. We show how cash holdings and returns affect the returns of standard asset pricing strategies and asset pricing models like the capital asset pricing model (CAPM).

As cash holdings of public firms increase, it is important that policymakers understand how these increases impact stock returns for both individual firms and the aggregate market. Cash returns are correlated across firms, and cash biases the measurement of the interconnectedness of stock returns. This correlation is important both for investors who are managing a portfolio’s risk and policymakers concerned about sources of vulnerability stemming from interconnected returns.

Credit Conditions in the Pandemic Mortgage Market

[from the Federal Reserve Bank of San Francisco]

by John Mondragon

The recent rapid rise in house prices has raised some questions about the potential risk to broader financial stability. However, credit quality in the mortgage market appears to be very high, and lending standards tightened in early 2020. While low interest rates increased the demand for refinancing, evidence from large nonconforming loans shows that credit supply contracted sharply in March 2020 and remained tight through the early pandemic period. The shift in credit supply suggests that lenders adjusted their standards to mitigate some risk in the housing market.

Read the full article [Archived PDF]