Working Papers

Using new data from the Understanding Society: COVID 19 survey collected in April 2020, we show how the aggregate shock caused by the pandemic affects individuals across the distribution. The survey collects data from existing members of the Understanding Society panel survey who have been followed for up to 10 years. Understanding society is based on probability samples and the Understanding Society Covid19 Survey is carefully constructed to support valid population inferences. Further the panel allows comparisons with a pre-pandemic baseline. We document how the shock of the pandemic translates into different economic shocks for different types of worker: those with less education and precarious employment face the biggest economic shocks.Some of those affected are able to mitigate the impact of the economic shocks: universal credit protects those in the bottom quintile, for example. We estimate the prevalence of the different measures individuals and households take to mitigate the shocks. We show that the opportunities for mitigation are most limited for those most in need.

JEL Codes: C83, D31, G51, I31, J31, J63

Keywords: COVID-19, job loss, inequality, mitigation, financial distress

Reference: 911

Individual View

In combating the spread of COVID-19, some governments have been reluctant to adopt lockdown policies due to their perceived economic costs. Such costs can, however, arise even in the absence of restrictive policies, if individuals’ independent reaction to the virus slows down the economy. This paper finds that imposing lockdowns leads to lower overall costs to the economy than staying open. We combine detailed location trace data from 40 million mobile devices with difference-in-differences estimations and a modification of the epidemiological SIR model that allows for societal and political response to the virus. In that way, we show that voluntary reaction incurs substantial economic costs, while the additional economic costs arising from lockdown policies are small compared to their large benefits in terms of reduced medical costs. Our results hold for practically all realistic estimates of lockdown efficiency and voluntary response strength. We quantify the counterfactual costs of voluntary social distancing for various US states that implemented lockdowns. For the US as a whole, we estimate that lockdowns reduce the costs of the pandemic by 1.7% of annual GDP per capita, compared to purely voluntary responses.

JEL Codes: I12, I18, H12, D04, C33, H51

Keywords: COVID-19, difference-in-differences, SIR model, social distancing, lockdown, big data

Reference: 910

Individual View

We develop tests for out-of-sample forecast comparisons based on loss functions that contain shape parameters. Examples include comparisons using average utility across a range of values for the level of risk aversion, comparisons of forecast accuracy using characteristics of a portfolio return across a range of values for the portfolio weight vector, and comparisons using a recently-proposed “Murphy diagrams” for classes of consistent scoring rules. An extensive Monte Carlo study verifies that our tests have good size and power properties in realistic sample sizes, particularly when compared with existing methods which break down when then number of values considered for the shape parameter grows. We present three empirical illustrations of the new test.

JEL Codes: C53, C52, C12

Keywords: Forecasting, model selection, out-of-sample testing, nuisance parameters

Reference: 909

Individual View

This paper analyses theoretically and quantitatively the effect that different higher education funding policies have on welfare (on aggregate and at the individual level) and wealth inequality. A heterogeneous agent model in continuous time, which has uninsurable income risk and endogenous educational choice is used to evaluate five different higher education financing schemes. Educational investments can be self financed, supported by government guaranteed student loans - that may come with or without income contingent support - or be covered by the public sector. When educational costs are small, differences in outcomes amongst systems are negligible. On the other hand, when these costs rise to realistic levels we see that there can be large gains in welfare and significant drops in inequality by moving to a system with more public sector support. This support can come in the form of tuition subsidies and/or income contingent student loans. However, as the cost of education and the share of debtors in society gets larger, it is preferable to increase public support in the form of tuition subsidies. The reason is that there is a pecuniary externality of debt that gets magnified when student loans become excessive. While I identify large steady state welfare gains from more public sector financing, I show that the transition costs can be large enough to justify the status quo.

JEL Codes: D52, D58, E24, I22, I23

Keywords: Incomplete markets, Higher education funding, Human capital

Reference: 908

Individual View

We show how to efficiently use costly testing resources in an epidemic, when testing outcomes can be used to make quarantine decisions. If the cost of false quarantine and false release exceed the cost of testing, the optimal myopic testing policy targets individuals with an intermediate likelihood of being infected. A high cost of false release means that testing is optimal for individuals with a low probability of infection, and a high cost of false quarantine means that testing is optimal for individuals with a high probability of infection. If individuals arrive over time, the policy-maker faces a dynamic tradeoff: using tests for individuals for whom testing yields the maximum immediate benefit vs. spreading out testing capacity across the population to learn prevalence rates thereby benefiting later individuals. We describe a simple policy that is nearly optimal from a dynamic perspective. We briefly discuss practical aspects of implementing our proposed policy, including imperfect testing technology, appropriate choice of prior, and non-stationarity of the prevalence rate.

Reference: 907

Individual View

How did countries recover from the Great Depression? In this paper we explore the argument that leaving the gold standard helped by boosting inflationary expectations and lowering real interest rates. We do so for a sample of 30 countries, using modern nowcasting methods and a new dataset containing more than 230,000 monthly and quarterly observations for over 1,500 variables. In those cases where the departure from gold happened on clearly defined dates, it seems clear that inflationary expectations rose in the wake of departure. Synthetic matching techniques suggest that the relationship is causal.

Revised May 2020.

Reference: 896

Individual View

Abstract

We evaluate the responsiveness of migrant self-selection to short-run changes in theWe evaluate the responsiveness of migrant self-selection to short-run changes in theeconomic environment. Using novel historical micro data, we estimate the initialselectivity of Mexican migration (1906-08) and focus on labor institutions as short-runadjustment channels of self-selection. We find that the first Mexican migrants werepositively self-selected on the basis of height—a proxy for physical productivity oflabor. Additionally, the US financial crisis of 1907 significantly modified self-selection.Shifts in migrant self-selection during and after the crisis were influenced by theenganche, an institution that reduced migration costs, but only for the “best” Mexicansduring “good” economic times.

JEL Codes: F22, J61, N36, O15

Keywords: labor migration, migrant self-selection, Panic of 1907, Mexico

Reference: 179

Individual View

The ratio estimator of a firm’s markup is the ratio of the output elasticity ofThe ratio estimator of a firm’s markup is the ratio of the output elasticity ofa variable input to that input’s cost share in revenue. This note raises issues thatconcern identification and estimation of markups using the ratio estimator. Concerningidentification: (i) if the revenue elasticity is used in place of the output elasticity, thenthe estimand underlying the ratio estimator does not contain any information aboutthe markup; (ii) if any part of the input bundle is either used to influence demand, or isneither fully fixed nor fully flexible, then the estimand underlying the ratio estimatoris not equal to the markup. Concerning estimation: (i) even with data on outputquantities, it is challenging to obtain consistent estimates of output elasticities whenfirms have market power; (ii) without data on output quantities, as is typically thecase, it is not possible to obtain consistent estimates of output elasticities when firmshave market power and markups are heterogeneous. These issues cast doubt overwhether anything useful can be learned about heterogeneity or trends in markups,from recent attempts to apply the ratio estimator in settings without output quantitydata.

JEL Codes: D2, D4, L1, L2

Keywords: Markups, Output Elasticity, Revenue Elasticity, Production Functions

Reference: 906

Individual View

Abstract: Seeking substantive relationships among vast numbers of spurious connections when modelling Big Data requires an appropriate approach. Big Data are useful if they can increase the probability that the data generation process is nested in the postulated model, increase the power of specification and mis-specification tests, and yet do not raise the chances of adventitious significance. Simply choosing the best-fitting equation or trying hundreds of empirical fits and selecting a preferred one–perhaps contradicted by others that go unreported–is not going to lead to a useful outcome. Wide-sense non-stationarity (including both distributional shifts and integrated data) must be taken into account. The paper discusses the use of principal components analysis to identify cointegrating relations as a route to handling that aspect of non-stationary big data, along with saturation to handle distributional shifts, and models the monthly UK unemployment rate, using both macroeconomic and Google Trends data, searching over 3000 explanatory variables and yet identifying a parsimonious, well-specified and theoretically interpretable model specification.

JEL Codes: C51, Q54

Keywords: Cointegration; Big Data; Model Selection; Outliers; Indicator Saturation; Autometrics

Reference: 905

Individual View

We study a canonical model of simultaneous price competition between firms thatWe study a canonical model of simultaneous price competition between firms thatsell a homogeneous good to consumers who are characterized by the number ofprices they are exogenously aware of. Our setting subsumes many employed in theliterature over the last several decades. We show there is a unique equilibrium ifand only if there exist some consumers who are aware of exactly two prices. Theequilibrium we derive is in symmetric mixed strategies. Furthermore, when thereare no consumers aware of exactly two prices, we show there is an uncountableinfinityof asymmetric equilibria in addition to the symmetric equilibrium. Ourresults show the paradigm generically produces a unique equilibrium. We also showthat the commonly-sought symmetric equilibrium (which also nests the textbookBertrand pure strategy equilibrium as a special case) is robust to perturbations inconsumer behavior, while the asymmetric equilibria are not.

Revised March 2020.

JEL Codes: D43, L11

Keywords: price competition; price dispersion; unique equilibrium

Reference: 874

Individual View

Abstract

We study strategic interactions in markets where rms sell to consumers both directlyWe study strategic interactions in markets where rms sell to consumers both directlyand via a competitive channel (CC), such as a price comparison website ormarketplace, where multiple sellers' oers are visible at once. We ask how a CC'ssize inuences market outcomes. A bigger CC means more consumers compareprices, increasing within-channel competition. However, we show such seeminglypro-competitive developments can raise prices and harm consumers by weakeningbetween-channel competition. We also use the model to study relevant active policyissues including price clauses, integrated ownership structures, and access toconsumers' purchase data.

Revised April 2020.

JEL Codes: JEL D43, D83, L11, M3

Reference: 843

Individual View

Gravity as both fact and theory is one of the great success stories of recent researchGravity as both fact and theory is one of the great success stories of recent researchon international trade, and has featured prominently in the policy debate overBrexit. We rst review the facts, noting the overwhelming evidence that trade tendsto fall with distance. We then introduce some expository tools for understanding CEStheories of gravity as a simple general-equilibrium system. Next, we point out someanomalies with the theory: mounting evidence against constant trade elasticities, and implausible predictions for bilateral trade balances. Finally, we sketch an approachimplausible predictions for bilateral trade balances. Finally, we sketch an approachbased on subconvex gravity as a promising direction to resolving them.

JEL Codes: F17, F14, F10

Keywords: Bilateral Trade Balances; Brexit; Elasticity of Trade to Distance; Quantile Regression; Structural Gravity and Trade; Subconvex Demands

Reference: 904

Individual View

Economic adjustment to trade and policy shocks is hampered by the fact that some sectors tend to cluster, so are hard to initiate in new places. This can give rise to persistent spatial disparities between cities within a country. The paper sets out a two-sector model in which cities divide into those producing tradable goods or services subject to agglomeration economies, and those only producing non-tradables for the national market. If import competition destroys some established tradable sectors, then affected cities fail to attract new tradable activities and switch to just produce non-tradables. Full employment is maintained (we assume perfect markets and price flexibility) but disparities between the two types of cities are increased. All non-tradable cities experience real income loss, while remaining tradable cities boom. The main beneficiaries are land-owners in remaining tradable cities, but there may be aggregate loss as the country ends up with too many cities producing non-tradables, and too few with internationally competitive activities. Fiscal policy has opposite effects in the two types of cities, with fiscal contraction causing decline in cities producing non-tradables, increasing activity in cities producing tradable goods, widening spatial disparities, and in the process increasing the share of rent in the economy.

JEL Codes: F12, F60, R11, R12

Keywords: Urban economics, divergence, rebalancing, lagging regions, de-industrialisation, fiscal policy

Reference: 903

Individual View

This paper documents state dependence in labor market fluctuations. Using a Threshold Vector Autoregression model (TVAR), we establish that the unemployment rate, the job separation rate, and the job finding rate exhibit a larger response to productivity shocks during periods with low aggregate productivity. A Diamond-Mortensen-Pissarides model with endogenous job separation and on-the-job search replicates these empirical regularities well. We calibrate the model to match the standard deviation of the job-transition rates explained by productivity shocks in the TVAR, and show that the model explains 88 percent of the state dependence in the unemployment rate, 76 percent for the separation rate and 36 percent for the job finding rate. The key channel underpinning state dependence in both job separation and job finding rates is the interaction of the firm’s reservation productivity level and the distribution of match-specific idiosyncratic productivity. Results are robust across several variations to the baseline model.

JEL Codes: E24, E32, J64, C11

Keywords: Search and Matching Models, State Dependence in Business Cycles, Threshold Vector Autoregression

Reference: 902

Individual View

A problem for both historical and contemporary research on inequality is a scarcity of high quality data on wealthy households. In this paper we explore a rich source of such data for historical periods: the account books of aristocratic households preserved in their family archives. We make three contributions: i) a survey of the nobility in Italy and of their publicly accessible archives; ii) an assay of the type and quality of budget data they contain; and iii) an assessment of the impact of adding upper-tail families to a household budget sample on inequality estimates. In a nutshell, our assessment is that the data are relatively abundant, accurate, and highly impactful. An enhanced sample of noble families will enable us to significantly improve estimates of Italian inequality right back to the country’s founding in 1861. There is no reason to think the approach would be any less feasible or fruitful in other European countries.

JEL Codes: N33, N34, I3

Reference: 178

Individual View

A century after Mussolini’s seizure of power, distributive trends during Interwar Italy are only partially known. This paper presents new evidence on inequality, contributing to the ‘classic’ debate on Fascism’s origins and legacy. Labour shares fell dramatically during the Great War, quickly recovered by 1922, and experienced a steady decline during Fascism, reaching a secular minimum in early 1940s. A newly assembled database of fiscal tabulations shows increasing concentration at the top between 1925 and 1936. These findings testify the fundamentally regressive nature of the Fascist regime, revealing significant discontinuity in Italy’s long-run inequality trend.

JEL Codes: B12, D63, J31, N14, N34

Reference: 177

Individual View

Social norms are costly if they are harmful for individuals but they remain in place for long periods of time because deviations are punished by members of the community. Examples include female genital cutting, foot binding, and codes of honour such as duelling. These and many other costly norms are seldom ‘all or nothing’: they are multidimensional and can take many altern¬ative forms. We develop a general theory of norm dynamics that focuses on the intermediate-run behaviour of such systems. Al-though in the (very) long run costly norms tend to die out, in the intermediate run transitions to less costly versions of the norm may occur that significantly retard its ultimate abandonment.

Revised January 2020.

Reference: 883

Individual View

Asset pricing and climate policy are analyzed in a global economy where consumption goods are produced by both a green and a carbon-intensive (dirty) sector. We allow for endogenous growth and three types of damages from global warming. It is shown that, initially, the desire to diversify assets in the portfolio complements the attempt to mitigate economic damages from climate change. In the long run, however, there is a trade-off between diversification and climate action. Therefore, in general, the carbon-intensive sector is not shut down completely. We derive the optimal carbon price, the equilibrium risk-free rate, and the risk premium of both assets. The risk-free rate is negatively affected by temperature, while the effect of temperature on the risk premiums depends on the type of damage specification. Climate disasters with an uncertain timing that rises with on temperature leads to a significant effect of climate change on asset prices.

JEL Codes: D81, G01, G12, Q5, Q54

Keywords: climate finance, decarbonization, diversification, carbon price, asset prices, green assets, disaster risk

Reference: 901

Individual View

We show that several of the most important economic models of climate change produce climate dynamics inconsistent with the current crop of models in climate science. First, most economic models exhibit far too long a delay between an impulse of CO2 emissions and warm¬ing. Second, few economic models incorporate positive feedbacks in the carbon cycle, whereby carbon sinks remove less CO2 from the atmosphere, the more CO2 they have already removed cumulatively, and the higher is temperature. These inconsistencies affect economic prescriptions to abate CO2 emissions. Controlling for how the economy is represented, different climate mod¬els result in significantly different optimal CO2 emissions. A long delay between emissions and warming leads to optimal carbon prices that are too low and too much sensitivity of optimal carbon prices to the discount rate. Omitting positive carbon cycle feedbacks also leads to op¬timal carbon prices that are too low. We conclude it is important for policy purposes to bring economic models in line with the state of the art in climate science.

JEL Codes: Q54

Keywords: carbon cycle, carbon price, climate change, integrated assessment modelling, positive feedbacks, social cost of carbon

Reference: 900

Individual View

With economic development can come social, attitudinal and cultural change, for good or ill or both We pose an unexplored question: why has happiness fallen in rural China whereas rural income has risen rapidly? Two rich data sets are analysed, the rural surveys of the China Household Income Project (CHIP) relating to 2002 and 2013. Our main methods are happiness regressions and decomposition methodology. Several approaches are adopted and no fewer than ten hypotheses are tested. One approach is to examine the variables that are found to be important in happiness functions and to consider their contributions to the fall in the mean happiness score of rural people. Another approach is to analyse the effect on rural happiness of the vast rural-urban migration that took place over this period. This is followed up by introducing tests of the role that changing attitudes might have played.

Reference: 899

Individual View

We investigate past climate variability over the Ice Ages, where a simultaneous-equations system is developed to characterize land ice volume, temperature and atmospheric CO2 levels as non-linear functions of measures of the Earth’s orbital path round the Sun. Although the orbital variables were first theorised as the fundamental causes of glacial variation by Croll in 1875 following Agassiz’s conception of a ‘Great Ice Age’ in 1840, their minor variations were thought insufficient to drive such major changes, especially the relative rapidity of shifts between glacial and warmer periods. The changes over the ice ages in atmospheric CO2 closely matched changes in land ice volumes, and since temperature changes are in turn affected by CO2 and also closely tracked ice volumes, a key identification issue is the causal role of CO2 in the process. As any links between CO2 and temperature above the forces from the orbital drivers (which of course are still operating) must have been natural ones hundreds of thousands of years ago, understanding their interactions at that time is important now that additional CO2 emissions are anthropogenic. We develop a simultaneous equation system over the last 800,000 years that allows a test of the role of CO2 as endogenously driven by the orbital variations, or an ‘exogenous’ influence as it now is.

JEL Codes: C01, C51, C87, Q54

Keywords: Climate Econometrics; Model Selection; Outliers; Identification; Saturation Estimation; Au- tometrics; Ice Ages

Reference: 898

Individual View

This paper develops a computable spatial equilibrium model of Central Asia and uses it to analyze the possible effects of the Belt and Road Initiative on the economy of the region. The model captures international and subnational economic units and their connectivity to each other and the rest of the world. Aggregate real income gains from the Belt Road Initiative range from less than 2 percent of regional income if adjustment mechanisms take the form of conventional Armington and monopolistic competition, to around 3 percent if there are localization economies of scale and labor mobility. In the latter case, there are sizeable geographical variations in impact, with some areas developing clusters of economic activity with income increases of as much as 12 percent and a doubling of local populations, while other areas stagnate or even decline.

JEL Codes: F12, F15, R11, R13

Keywords: regional integration, transport infrastructure, spatial modeling, economic geography, Central Asia.

Reference: 897

Individual View

The recent international financial crisis has dramatically revealed the shortcomings and potential dangers of bank globalization and deeper financial integration. While the lack of supervision and adequate legal frameworks has been largely acknowledged as a main problem, the effects of regulation on the development of international banking activity and financial stability are still a matter of controversy. This article investigates the conditions under which the globalization of the domestic banking sector unfolded in Brazil and Mexico during the years of dizzying expansion of foreign finance that culminated the international debt crisis of 1982. It shows how the regulatory framework for international banking and foreign capital in Brazil created a model of intermediation that was considerably less vulnerable to crisis than in Mexico, with a more lightly regulated institutional base. These findings provide insights into historical discussions about the implications of financial regulation and capital controls for the development and expansion of foreign finance and whether the risks underlying international banking are necessarily inherent to the process of financial globalization.

Reference: 176

Individual View

This paper uses the first discovery of oil a nd gas in Norway as a natural experiment to study the long run labour market implications of a positive economic shock. Existing studies largely focus on short term dynamics and on men, but the effects on women and their persistence in time are less known. Following the same individuals for up to two decades in the Norwegian Registry and Census data, we find that the oil discovery significantly increased male earnings (up to 7% annually), while female earnings declined (more than 10%). The shift in annual income is still present 15 years after the discovery. Labour force participation increased among men, while it declined for women in full-time employment. Specifically, the decline in female earnings was driven by married women. Moreover, men in oil regions shifted into high-paying occupations while women did the opposite after the discovery of oil. However, the income loss for women is mostly a short-term phenomenon: we find that women’s lifetime income improved for later cohorts.

Reference: 223

Individual View

Can oil discovery shocks affect the demand for protectionism? A two-period model of Dutch disease indicates that if the tradable sector is politically dominant then an oil discovery induces protectionism. If the economy is also credit constrained, this effect is intensified upon discovery, but partially reversed when oil revenues start to flow. We test these predictions using detailed bilateral tariff data that cover 96 products in 155 countries over the period 1988-2012, and worldwide discoveries of giant oil and gas fields. Our identification strategy rests on the exogeneity of the timing of discoveries. We find that an oil discovery increases tariffs during pre-production years and decreases tariffs in the years to follow yet to a lesser extent, most notably in capital scarce economies with a relatively dominant tradable sector. Our baseline estimates indicate that a giant oil field discovery induces a rise of approximately 15% in the average tariff over the course of 10 years; this increase is about 1.8 times larger during the pre¬production period when the oil discovery represents a pure news shock.

JEL Codes: Q32, F13, O24

Keywords: Oil discoveries, protectionism, capital scarcity, Dutch disease, political economy, trade policy, news shocks

Reference: 895

Individual View