We provide a critical review of macroeconomic models used for monetary policy at central banks from a finance perspective. We review the history of monetary policy modeling, survey the core monetary models used by major central banks, and construct an illustrative model for those readers who are unfamiliar with the literature. Within this framework, we highlight several important limitations of current models and methods, including the fact that local-linearization approximations omit important nonlinear dynamics, yielding biased impulse-response analysis and parameter estimates. We also propose new features for the next generation of macrofinancial policy models, including: a substantial role for a financial sector, the government balance sheet and unconventional monetary policies; heterogeneity, reallocation, and redistribution effects; the macroeconomic impact of large nonlinear risk-premium dynamics; time-varying uncertainty; financial sector and systemic risks; imperfect product market and markups; and further advances in solution, estimation, and evaluation methods for dynamic quantitative structural models.

We provide a review of macro-finance models featuring nonlinear dynamics. We survey the models developed recently in the literature, including models with amplification effects of financial constraints, models with households' leverage constraints, and models with financial networks. We also construct an illustrative model for those readers who are unfamiliar with the literature. Within this framework, we highlight several important limitations of local solution methods compared with global solution methods, including the fact that local-linearization approximations omit important nonlinear dynamics, yielding biased impulse-response analysis.

We represent risk factors as sums of orthogonal components capturing fluctuations with cycles of different length. The representation leads to novel spectral factor models in which systematic risk is allowed (without being forced) to vary across frequencies. Frequency-specific systematic risk is captured by a notion of spectral beta. We show that traditional factor models restrict the spectral betas to be constant over frequencies. The restriction can hide horizon-specific pricing effects which spectral factor models are designed to reveal. We illustrate how the methods may lead to economically-meaningful dimensionality reduction in the factor space.

Advances in financial technology have made tax-loss harvesting more feasible for retail investors than such strategies were in the past. We evaluated the magnitude of this “tax alpha” with the use of historical data from the CRSP monthly database for the 500 securities with the largest market capitalizations from 1926 to 2018. Given long-term and short-term capital gains tax rates of 15% and 35%, respectively, we found that a tax-loss-harvesting strategy yielded a before-transaction-cost tax alpha of 1.08% per year for our sample period. When the strategy was constrained by the “wash sale rule,” the tax alpha decreased from 1.08% per year to 0.82% per year.

In the midst of epidemics such as COVID-19, therapeutic candidates are unlikely to be able to complete the usual multi-year clinical trial and regulatory approval process within the course of an outbreak. We apply a Bayesian adaptive patient-centered model—which minimizes the expected harm of false positives and false negatives—to optimize the clinical trial development path during such outbreaks. When the epidemic is more infectious and fatal, the Bayesian-optimal sample size in the clinical trial is lower and the optimal statistical significance level is higher. For COVID-19 (assuming a static Ro = 2 and initial infection percentage of 0.1%), the optimal significance level is 7.1% for a clinical trial of a non-vaccine anti-infective therapeutic clinical trial and 13.6% for that of a vaccine. For a dynamic Ro ranging from 2 to 4, the corresponding values are 14.4% and 26.4%, respectively. Our results illustrate the importance of adapting the clinical trial design and the regulatory approval process to the specific parameters and stage of the epidemic.

A key driver in biopharmaceutical investment decisions is the probability of success of a drug development program. We estimate the probabilities of success (PoS) of clinical trials for vaccines and other anti-infective therapeutics using 43,414 unique triplets of clinical trial, drug, and disease between January 1, 2000, and January 7, 2020, yielding 2,544 vaccine programs and 6,829 non-vaccine programs targeting infectious diseases. The overall estimated PoS for an industry-sponsored vaccine program is 39.6%, and 16.3% for an industry-sponsored anti-infective therapeutic. Among industry-sponsored vaccines programs, only 12 out of 27 disease categories have seen at least one approval, with the most successful being against monkeypox (100%), rotavirus (78.7%), and Japanese encephalitis (67.6%). The three infectious diseases with the highest PoS for industry-sponsored nonvaccine therapeutics are smallpox (100%), CMV (31.8%), and onychomycosis (29.8%). Nonindustry- sponsored vaccine and non-vaccine development programs have lower overall PoSs: 6.8% and 8.2%, respectively. Viruses involved in recent outbreaks—MERS, SARS, Ebola, Zika—have had a combined total of only 45 non-vaccine development programs initiated over the past two decades, and no approved therapy to date (Note: our data was obtained just before the COVID-19 outbreak and do not contain information about the programs targeting this disease.) These estimates offer guidance both to biopharma investors as well as to policymakers seeking to identify areas most likely to be undeserved by private-sector engagement and in need of public-sector support.

The healthcare industry in the United States (U.S.) is a complex ecosystem with many different stakeholders. Unlike the universal single-payer healthcare systems of many European countries,the accessibility of prescription drugs in the U.S. is largely determined by contract negotiations between health plans and drug manufacturers about formulary placement. These negotiations can sometimes result in higher out-of-pocket costs for the patient, since the current structure of the U.S. healthcare system creates a perverse incentive for many health plans to elicit higher rebates from drug manufacturers in exchange for formulary placement of brand-name drugs, thereby increasing patients’ out-of-pocket costs.

We define long shots as investment projects with four features: (1) low probabilities of success; (2) long gestation lags before any cash flows are realized; (3) large required up-front investments; and (4) very large payoffs (relative to initial investment) in the unlikely event of success. Funding long shots is becoming increasingly difficult—even for high-risk investment vehicles like hedge funds and venture funds—despite the fact that some of society’s biggest challenges such as cancer, Alzheimer’s disease, global warming, and fossil-fuel depletion depend critically on the ability to undertake such investments. We investigate the possibility of improving financing for long shots by pooling them into a single portfolio that can be financed via securitized debt, and examine the conditions under which such funding mechanisms are likely to be effective.

We extend earlier large-scale studies of clinical trial statistics by focusing on the performance of oncology trials. Using 108,248 data points between January 1, 2005, and September 31, 2018, compiled from the Citeline database, we investigate the duration of clinical trials and compute the probabilities of success of 24,448 oncology drug development programs by disease group. While the overall phase 1 to approval rate for all oncology-related drug development programs is 3.3%, individual disease groups have approval rates ranging from 0% to 10.1%. Similar patterns can be seen for oncology orphan drug development programs, where the overall probability of success ranges from 0% to 8.3%, with an overall average of 1.9%. We find overwhelming evidence that using biomarkers for patient selection is effective in almost all disease groups within oncology, raising the overall probability of success by an average of 13.3%.

Advances in biomedical research have created significant opportunities to bring to market a new generation of therapeutics. However, early-stage assets often face a dearth of funding, as they have a high risk of failure and significant development costs. Historically, this has been particularly true for assets intended to treat rare diseases, where market sizes are often too small to attract much attention and funding. Venture philanthropy (VP) — which, for the purpose of this paper, is defined as a model in which nonprofit, mission-driven organizations fund initiatives to advance their objectives and potentially achieve returns that can be reinvested toward their mission — offers an alternative to traditional funding sources like venture capital or the public markets. Here we highlight the Cystic Fibrosis (CF) Foundation, widely considered to be the leading VP organization in biotech, which facilitated the development of Kalydeco, the first disease-modifying therapy approved to treat cystic fibrosis. We evaluate the CF Foundation’s example, including its agreement structures and strategy, explore the challenges that other nonprofits may have in adopting this strategy, and draw lessons from the CF Foundation for other applications of VP financing.

The prediction of clinical outcomes for patients with cancer is central to precision medicine and the design of clinical trials. We developed and validated machine-learning models for three important clinical end points in patients with advanced non–small-cell lung cancer (NSCLC)—objective response (OR), progression free survival (PFS), and overall survival (OS)—using routinely collected patient and disease variables. We aggregated patient-level data from 17 randomized clinical trials recently submitted to the US Food and Drug Administration evaluating molecularly targeted therapy and immunotherapy in patients with advanced NSCLC. To our knowledge, this is one of the largest studies of NSCLC to consider biomarker and inhibitor therapy as candidate predictive variables. We developed a stochastic tumor growth model to predict tumor response and explored the performance of a range of machine-learning algorithms and survival models. Models were evaluated on out-of-sample data using the standard area under the receiver operating characteristic curve and concordance index (C-index) performance metrics. Our models achieved promising out-of-sample predictive performances of 0.79 area under the receiver operating characteristic curve (95% CI, 0.77 to 0.81), 0.67 C-index (95% CI, 0.66 to 0.69), and 0.73 C-index (95% CI, 0.72 to 0.74) for OR, PFS, and OS, respectively. The calibration plots for PFS and OS suggested good agreement between actual and predicted survival probabilities. In addition, the Kaplan-Meier survival curves showed that the difference in survival between the low- and high-risk groups was significant (log-rank test P, .001) for both PFS and OS. Biomarker status was the strongest predictor of OR, PFS, and OS in patients with advanced NSCLC treated with immune checkpoint inhibitors and targeted therapies. However, single biomarkers have limited predictive value, especially for programmed death-ligand 1 immunotherapy. To advance beyond the results achieved in this study, more comprehensive data on composite multiomic signatures is required.

A commentary on the article, "Artificial Intelligence—The Revolution Hasn’t Happened Yet" by Michael I. Jordan, published by Harvard Data Science Review (July 2019).

Researchers, clinicians, policymakers and patients are increasingly interested in questions about therapeutic interventions that are difficult or costly to answer with traditional, free-standing, parallel-group randomized controlled trials (RCTs). Examples include scenarios in which there is a desire to compare multiple interventions, to generate separate effect estimates across subgroups of patients with distinct but related conditions or clinical features, or to minimize downtime between trials. In response, researchers have proposed new RCT designs such as adaptive platform trials (APTs), which are able to study multiple interventions in a disease or condition in a perpetual manner, with interventions entering and leaving the platform on the basis of a predefined decision algorithm. APTs offer innovations that could reshape clinical trials, and several APTs are now funded in various disease areas. With the aim of facilitating the use of APTs, here we review common features and issues that arise with such trials, and offer recommendations to promote best practices in their design, conduct, oversight and reporting.

Billions of dollars are spent annually on cancer drug development, yet effective treatments for many types of cancer remain as elusive as ever. Recently, the MIT Laboratory for Financial Engineering announced the launch of Project ALPHA (Analytics for Life-sciences Professionals and Healthcare Advocates), a large-scale estimation of clinical trial probabilities of success (PoS) for a variety of drug development programs, where a single program is defined as the set of all clinical trials corresponding to a unique drug-indication pair. In that study, we found that only 3.4 percent of all cancer drug development programs from 2000 to 2015 moved from phase 1 to regulatory approval, despite the fact that oncology accounted for 42 percent of all drug development programs in that dataset.