Michaela Pagel

Publications: 

Retirement Puzzles: New Evidence from Personal Finances (joint with Arna Olafsson, accepted, Journal of Public Economics, 2022) click Using a comprehensive transaction-level panel data set, we document that individuals repay their consumer debt and save more after they retire. These findings are unexpected because people should save before the fall in income at retirement, rather than starting to save after. We carefully discuss a number of explanations for our findings, including a drop in work-related expenses and an increase in medical health risks around retirement. These two mechanisms are the leading explanations of the so-called retirement-consumption and retirement-savings puzzles, which allows our findings to inform the larger question of whether individuals save enough for retirement. Additionally, we rationalize our findings in a model with non-standard preferences.

Media coverage: VoxEU, MarketWatch

Selected presentations: Federal Reserve Richmond, Frankfurt School of Finance,  CEPR European Conference on Household Finance, PerCent 2017 Conference, Minnesota Junior Conference, ECWFC at the WFA 2017, AFA, nominated for the Distinguished CESifo Affiliate Award at the 8th conference of the CESifo network on Behavioral Economics, NBER Behavioral Macro Summer Institute, CEPRA/NBER Workshop on Ageing and Health

Fresh Air Eases Work – The Effect of Air Quality on Individual Investor Activity (joint with Steffen Meyer, accepted, Review of Finance, 2023) click This paper shows that air quality has a significantly negative effect on the likelihood of individual investors to sit down, log in, and trade in their brokerage accounts controlling for investor-, weather-, traffic-, and market-specific factors. In perspective, a one standard deviation increase in fine particulate matter leads to a 9% reduction in the probability of logging in and trading, which is larger than the reduction due to a one standard deviation increase in sunshine. As individual investor trading can be a proxy for everyday cognitively-demanding tasks such as office work, our findings suggest that the negative effects of pollution on white-collar work productivity are much more severe than previously thought for levels of pollution that are commonly found throughout the developed world. To our knowledge, this is the first study to demonstrate a negative impact of pollution on a measure of white-collar productivity at the individual level in a Western country.

Media coverage: Washington Post, The Morning Call, World Economic Forum, LA Times

Selected presentations: Columbia University financial economics workshop, SITE New Thinking about Economic Challenges in the Design and Implementation of Programs to Stabilize the Climate

Income, Liquidity, and the Consumption Response to the 2020 Economic Stimulus Payments (joint with Scott Baker, Robert Farrokhnia, Steffen Meyer, and Constantine Yannelis, Review of Finance, 2021) click The 2020 CARES Act directed large cash payments to households. We analyze households' spending responses using high-frequency transaction data from a FinTech nonprofit, exploring heterogeneity by income levels, recent income declines, and liquidity. Households respond rapidly to the receipt of stimulus payments, with spending increasing by $0.25-$0.30 per dollar of stimulus during the first weeks. Households with lower incomes, greater income drops, and lower levels of liquidity display stronger responses highlighting the importance of targeting. Liquidity plays the most important role, with no observed spending response for households with high levels of bank account balances. Relative to the effects of previous economic stimulus programs in 2001 and 2008, we see faster effects, smaller increases in durables spending, and larger increases in spending on food, likely reflecting the impact of shelter-in-place orders and supply disruptions. Additionally, we see substantial increases in payments like rents, mortgages, and credit cards reflecting a short-term debt overhang. We formally show that these differences can make direct payments less effective in stimulating aggregate consumption.

Selected presentations: Toulouse School of Finance, Barnard Applied Micro Lunch, Columbia Research Intern Lunch, Columbia finance lunch, RCFS/RAPS, Virtual AFFECT seminar, CEPR HF seminar, Philadelphia Fed, CEPR New Consumption Data Conference, USC, AEA, RAPS Bahamas Conference, Lenzerheide Ski Conference, 

Media coverage: The Economist, Chicago Booth Review, MarketWatch, Marketplace, Equitable Growth, CNN/ABC

Mobile Apps and Financial Decision Making (joint with Bruce Carlin and Arna Olafsson, Review of Finance, 2021) click We exploit the release of a mobile application for a financial aggregation platform to analyze how technology adoption changes consumer financial decision making. The app reduced the cost of accessing personal financial information, and we find that this led to a drop in non-sufficient fund (NSF) fees. Because of the manner in which these fees are incurred, this represents an unambiguous welfare improvement for users of the platform. The leading explanation for this result appears to be mistake avoidance due to easier access to information.

Media coverage: Ideas at Work, Think Forward Initiative

Selected presentations: AFFECT Conference University of Miami, University of Kentucky Finance Conference, Santiago Finance Conference, ITAM Finance Conference, AEA, RCFS Bahamas Conference, CERGE-EI Prague

Fully Closed: Individual Responses to Realized Gains and Losses (joint with Steffen Meyer, Journal of Finance, 2021) pdf code We use retail investor security trades and holdings data to analyze how individuals reinvest realized capital gains and losses. For identification, we exploit plausibly exogenous sales due to mutual fund liquidations. Theoretically, if individuals held optimized portfolios, we would expect them to simply reinvest everything out of the forced liquidations. Empirically, individuals reinvest – on average – 83% if the forced sale resulted in a gain relative to the initial investment, but only 40% in the event of a loss. This difference is statistically significant for more than six months and arises because many individuals forced to realize a loss choose to not reinvest at all and some even exit the stock market altogether. Our findings are consistent with mental accounting and the existing evidence that individuals treat realized losses differently from paper losses. Additionally, individuals may be scarred by experiencing losses in stock markets which might shed light on the stock market participation puzzle.

Media coverage: FinLit Interview 

Selected presentations: UCLA Anderson, University of Chicago Booth, Berkeley Haas, Harvard Economics, IESE, ECBE in Bergen Norway, NBER Household Finance Summer Institute, SITE Workshop Psychology and Economics, SFS Cavalcade, Boulder Summer Conference on Consumer Financial Decision Making, AFA, RAPS Bahamas Conference

Consumption Imputation Errors in Administrative Data (joint with Scott Baker, Lorenz Küng, and Steffen Meyer, Review of Financial Studies, 2021) pdf code Many research papers in household finance utilize annual snapshots of household wealth from administrative data, such as tax registries, to calculate "imputed consumption." However, trading costs, unobserved intra-year trades, or unobserved security characteristics may cause measurement error. We document how such errors vary across groups of individuals by income, portfolio characteristics, and wealth and how they are correlated with individual income and balance sheets, asset prices, and the business cycle using transaction-level retail brokerage account data. We find that the economic significance of imputation error is small in many research settings and we discuss robustness checks and econometric specifications to minimize the impact of imputation error in future research.

Selected presentations: NBER Trans-Atlantic Public Economics Seminar, NBER Summer Institute, Copenhagen Workshop on New Consumption Data

How Does Household Spending Respond to an Epidemic? Consumption During the 2020 COVID-19 Pandemic? (joint with Scott Baker, Robert Farrokhnia, Steffen Meyer, and Constantine Yannelis, Review of Asset Pricing Studies, Covid-19 and Financial Markets, 2020) pdf We explore how household consumption responds to epidemics, utilizing transaction-level household financial data to investigate the impact of the COVID-19 virus. As the number of cases grew, households began to radically alter their typical spending across a number of major categories. Initially spending increased sharply, particularly in retail, credit card spending and food items. This was followed by a sharp decrease in overall spending. Households responded most strongly in states with shelter-in-place orders in place by March 29th. We explore heterogeneity across partisan affiliation, demographics and income. Greater levels of social distancing are associated with drops in spending, particularly in restaurants and retail.

Media coverage: The Economist, Chicago Booth Review, MarketWatch, Marketplace, Equitable Growth

Sticking To Your Plan: Empirical Evidence on the Role of Present Bias for Credit Card Debt Paydown (joint with Theresa Kuchler, Journal of Financial Economics, 2019) pdf Using high-frequency transaction-level income, spending, balances, and credit limits data from an online financial service, we show that many consumers fail to stick to their self-set debt paydown plans and argue that this behavior is best explained by a model of present bias. Theoretically, we show that (i) a present-biased agent's sensitivity of consumption spending to paycheck receipt reflects his or her short-run impatience and that (ii) this sensitivity varies with available resources only for agents who are aware (sophisticated) rather than unaware (naive) of their future impatience. In turn, we classify users in our data accordingly. Consistent with present bias, we find that (i) sophisticated users' average paydown falls with higher measured impatience and that (ii) their planned paydown is more predictive of actual paydown than that of naives. We are the first to provide a theoretically-founded empirical methodology to measure naivete from spending and income data and validate this measure using our information on plans versus actual debt paydown. Moreover, our results highlight the importance of distinguishing between sophisticated and naive present-biased individuals in understanding their financial decision making.

Prospective Gain-Loss Utility: Ordered versus Separated Comparison (Journal of Economic Behavior & Organization, 2019) pdf code 

Generational Differences in Managing Personal Finances (joint with Bruce Carlin and Arna Olafsson, AEA Papers and Proceedings, 2019) pdf code

A News-Utility Theory for Inattention and Delegation in Portfolio Choice (Econometrica, 2018) pdf code Recent evidence suggests that investors are either inattentive to their portfolios or undertake puzzling rebalancing efforts. This paper develops a life-cycle portfolio-choice model in which the investor experiences loss-averse utility over news and can choose whether or not to look up his portfolio. I obtain three main predictions. First, the investor prefers to ignore and not rebalance his portfolio most of the time to avoid fluctuations in news utility. Such fluctuations cause a first-order decrease in expected utility because the investor dislikes bad news more than he likes good news. Consequently, the investor has a first-order willingness to pay a portfolio manager who rebalances actively on his behalf. Second, if the investor looks up his portfolio himself, he rebalances extensively to enjoy or delay the realization of good or bad news, respectively. Third, the investor would like to commit to being inattentive even more often because it reduces overconsumption. Quantitatively, I structurally estimate the preference parameters by matching participation and stock shares over the life cycle. My parameter estimates are in line with the literature, generate reasonable intervals of inattention, and simultaneously explain consumption and wealth accumulation over the life cycle. 

Media coverage: Ideas at Work, NY Times, US News, Tim Schaefer media, Think Advisor, CNN money, WirtschaftsBlatt, The Street, Finance No Free Lunch

The Liquid Hand-to-Mouth: Evidence from Personal Finance Management Software (joint with Arna Olafsson, Review of Financial Studies, 2018) pdf code We use a very accurate panel of individual spending, income, balances, and credit limits from a financial aggregation app and document significant payday responses of spending to the arrival of both regular and irregular income. These payday responses are clean, robust, and homogeneous for all income and spending categories throughout the income distribution. Spending responses to income are typically explained by households' financial structures: households that hold little or no liquid wealth have to consume hand-to-mouth. However, we find that few individuals hold little or no liquidity and also document that liquidity holdings are much larger than predicted by state-of-the-art models explaining spending responses with liquidity constraints due to illiquid savings. Given that present liquidity constraints do not bind, we analyze whether individuals hold cash cushions to cope with future liquidity constraints. To that end, we analyze cash holding responses to income payments inspired by the corporate finance literature. However, we find that individuals' cash responses are consistent with standard models without illiquid savings and neither present nor future liquidity constraints being frequently binding. Because these models are inconsistent with payday responses, we feel that the evidence suggests the existence of households that spend heuristically and call those the "liquid hand-to-mouth." 

Media coverage: Ideas at Work, Yahoo!, Finance, Wallethub, MSN Money, MoneyTalks News, Pblcty, Nordic Business Forum, Clark Howard, LostInEconLand, Investopedia, MyBankTracker

Expectations-Based Reference-Dependent Life-Cycle Consumption (Review of Economic Studies, 2017) pdf code This paper incorporates a recent preference specification of expectations-based loss aversion, which has been broadly applied in microeconomics, into a classic macro model to offer a unified explanation for three empirical observations about life-cycle consumption. First, loss aversion rationalizes excess smoothness and sensitivity, the empirical observation that consumption responds to income shocks with a lag. Intuitively, such lagged responses allow the agent to delay painful losses in consumption until his expectations have adjusted. Second, the preferences generate a hump-shaped consumption profile. Early in life, consumption is low due to a first-order precautionary-savings motive. But, as uncertainty resolves over time, this motive becomes dominated by time-inconsistent overconsumption that eventually leads to declining consumption toward the end of life. Third, consumption drops at retirement. Prior to retirement, the agent wants to overconsume his uncertain income before his expectations catch up. Post retirement, however, income is no longer uncertain, so that overconsumption is associated with a certain loss in future consumption. As an empirical contribution, I structurally estimate the preference parameters using life-cycle consumption data. My estimates match those obtained in experiments and other micro studies and generate the degree of excess smoothness observed in macro consumption data.

Expectations-Based Reference-Dependent Preferences and Asset Pricing (Journal of the European Economic Association, 2015) pdf code This paper incorporates expectations-based reference-dependent preferences into the canonical Lucas-tree asset-pricing economy. Expectations-based loss aversion increases the equity premium and decreases the consumption-wealth ratio, because uncertain fluctuations in consumption are more painful. Moreover, because unexpected cuts in consumption are particularly painful, the agent wants to postpone such cuts to let his reference point decrease. Thus, even though shocks are i.i.d., loss aversion induces variation in the consumption-wealth ratio, which generates variation in the equity premium, expected returns, and predictability. The level and variation in the equity premium and the predictability in returns match historical moments, but the associated variation in intertemporal substitution motives results in excessive variation in the risk-free rate. This effect can be partially offset with variation in expected consumption growth, heteroskedasticity in consumption growth, or time-variant disaster risk. As a key contribution, I show that the preferences resolve the equity-premium puzzle and simultaneously imply plausible risk attitudes towards small and large wealth bets beyond explaining microeconomic evidence in many other domains. 

Under review:

The Ostrich in Us: Selective Attention to Personal Finances (joint with Arna Olafsson, conditionally accepted, The Review of Economics and Statistics (ReStat), 2023) click A number of theoretical research papers across multiple fields in economics model and analyze attention but direct empirical evidence remains scarce. This paper investigates the determinants of attention to financial accounts using panel data from a financial management software provider containing daily logins, discretionary spending, income, balances, and credit limits. We first explore whether individuals pay attention in response to the arrival of income payments. Here, we utilize that weekends and holidays generate exogenous variation in regular payment arrival using a fixed-effects approach. We find that individuals are considerably more likely to log in because they get paid. Beyond looking at the causal effect of income on attention, we examine how attention depends on individual spending, balances, and credit limits relative to individuals' own histories. We find that attention is decreasing in spending and overdrafts and increasing in cash holdings, savings, and liquidity. Moreover, attention jumps discretely when balances change from negative to positive. We argue that all of our findings are consistent with Ostrich effects and anticipatory utility as the main motivation for paying attention to financial accounts and thus provide new tests for information- or belief-dependent models. Furthermore, we show that some of our findings can be explained by a recent influential one of those models assuming individuals experience utility over news or changes in expectations about consumption (Kőszegi and Rabin, 2009). 

Media coverage: VoxEU, Lifehacker

Selected presentations: Cornell, Maryland, 2017 BEAM at Berkeley, Carnegie Mellon, NBER Asset Pricing Meeting, NBER Digitization Summer Institute, University of Amsterdam, Jackson Hole Finance Conference, NAWFA, ECWFC at the WFA, EFA, AEA, NYU, TAU Finance, ESSFM Gerzensee, Zurich, Indiana University

Does Saving Cause Borrowing? Implications for the Co-Holding Puzzle (joint with Paolina Medina, revise and resubmit, Journal of Finance, 2022) click We analyze an experiment involving 3.1 million bank customers who were encouraged to save through SMS messages. We first theoretically show that by examining their spending, saving, and borrowing responses we can distinguish between the leading explanations for co-holding liquid savings and credit card debt. Using a machine learning algorithm, we then predict individual-level treatment effects and find that the most responsive individuals reduce spending and increase their savings by 5.1\% (225 USD PPP per month), while their credit card debt remains unchanged. We argue that these joint findings suggest people co-hold because they mentally separate savings and debt accounts. 

Selected presentations:  CEPR HF Conference, University of Regensburg, ANU, SAIF, Columbia Finance Lunch, ETH Zurich, CEPR Household Finance Seminar Series, Bank of Canada 2021 Annual Economics Conference on Behavioral Macro-Finance: Implications for Central Bankers (invited talk), CFPB research conference

Social Media coverage: CarlosHankGonzalez.com, Twitter

Working papers:

Consumption out of Fictitious Capital Gains and Selective Inattention (joint with Benjamin Loos and Steffen Meyer) click Do retail investors’ behavioral biases in trading directly affect their consumption out of stock market wealth? We exploit a natural experiment that changed the displayed purchase prices in investors’ online portfolios. Investors are more likely to sell and consume on average 25% of “fictitious” capital gains, i.e., displayed capital gains under the new purchase prices that are capital losses under the actual purchase prices. We argue that investors are selectively inattentive: they are more responsive when fictitious gains are larger and actual losses are smaller, they notice fictitious losses, and they react even when actual purchase prices are very salient..

Selected presentations: Columbia GSB Finance Lunch, Rising 5 Star Conference, 2nd Annual Conference for Women in Economics at Princeton, AQR Institute Academic Symposium, University of Mannheim, ESMT, University of Rotterdam, Endless Summer Conference, Copenhagen Workshop on New Consumption Data, CEU, University of Maastricht, Red Rock Conference, AEA, Northwestern Kellogg, CEU, EEA invited talk, CEPR European Workshop on Household Finance, WashU, LSE

Gender Quotas and Support for Women in Board Elections (joint with Marina Gertsberg and Johanna Mollerstrom) click We study shareholder support for corporate board nominees before and after the 2018 California gender quota. Pre-quota, new female nominees received greater support than new male nominees, consistent with women being held to a higher standard. Post-quota, as the number of women increased, support for new (mandated) female nominees decreased to the same level of support that new male nominees enjoy. Thus, shareholders approve of the quota-mandated female nominees. However, share prices reacted negatively to the quota. We show that this reaction was concentrated in firms that did not turn over their least-supported male directors when complying with the quota.

Selected presentations:  Barnard Applied Micro Lunch, ANU, Monash University, EFA, NBER-RFS Conference on Inequality, Discrimination and the Financial System, WFA, NBER Corporate Finance Meeting, European Winter Finance Conference 

Media coverage: Harvard Law School Forum on Corporate Governance, Daily Advent, Estadao, NBER Digest, Politico

The Effect of Stock Ownership on Individual Spending, Investments, and Loyalty (joint with Paolina Medina and Vrinda Mittal) click In this paper, we analyze how one of the most fundamental behavioral biases in investing---people's preference for buying specific stocks rather than holding the market portfolio---affects their life-cycle consumption, savings, and stock-market participation. We first show that when investors receive stocks from specific companies, they increase their spending in those companies' stores. While specific stock ownership increases total spending in the short run, individuals' overall stock market investments increase in the long run. For identification, we use the staggered allocation of brokerage accounts to individuals over time as well as quasi-randomly distributed stock grants.

Selected presentations:  CEPR Workshop on New Consumption Data, Columbia Finance Lunch, Georgetown Fintech Apps Seminar, McIntire University of Virginia, University of Maryland, University of Virginia, Hebrew University, WFA, EFA, NBER Household Finance Summer Institute, Cornell, CFPB research conference

Media coverage: Columbia GSB press release, TAMU press release, MarketWatch, Retail Bum, Vox

Consumer Surveillance and Financial Fraud (joint with Bo Bian, Devesh Raval, and Huan Tang) click In today's digital economy, firms near constantly collect, analyze, and profit from consumers' personal information, which might expose consumers to financial fraud. We examine the effects of Apple's App Tracking Transparency (ATT) policy, which significantly curtailed data collection and sharing on the iOS platform. Using zip code level variation in iOS user shares, we show that ATT substantially reduced fraud complaints. The effects are concentrated in complaints that have more relevant narratives and in complaints about companies engaging in intensive consumer surveillance and lacking data safeguards. Our evidence quantifies one of the main harms of lax privacy standards.

Selected presentations:  Microsoft Research, NBER SI (Household Finance), UBC Winter Finance Conference, University of Innsbruck, UCL, Junior Household Finance Brown Bag

Credit Smoothing (joint with Sean Hundtofte and Arna Olafsson) click Standard economic theory suggests that high-interest, unsecured, short-term borrowing, e.g., borrowing via credit cards, helps individuals smooth consumption in the event of transitory income shocks. This paper shows that, on average, individuals do not use such borrowing to smooth consumption when they experience a typical transitory income shock due to unemployment. Rather, it appears as if individuals smooth their roll-over credit card debt. We first use detailed longitudinal information on debit and credit account transactions, balances, and limits from a financial aggregator in Iceland to document that unemployment does not induce a large borrowing response at the individual level. We then replicate this finding in a representative sample of U.S. credit card holders, instrumenting local changes in employment using a Bartik-style instrument. The absence of a borrowing response occurs even when credit supply is ample and liquidity constraints do not bind (as captured by credit limits). This finding is difficult to reconcile with theories of consumption smoothing, which predict a strictly countercyclical demand for credit. On the contrary, the demand for credit does not appear to lean against business cycle fluctuations, leading to greater consumption volatility than what would be observed otherwise.

Media coverage: VoxEU Interview

Selected presentations:  NY Fed, Toulouse School of Economics, University of Mannheim, Nottingham University, University of St. Gallen, CEPR European Household Finance Conference, CFPB Research Conference, EEA, EMMM2, Colorado Finance Summit in Vail, AEA, MicroMacro Conference Chicago, WFA

Borrowing in Response to Windfalls (joint with Arna Olafsson) click We use high-accuracy and comprehensive transaction-level panel data containing information on all spending, income, balances, and credit limits of a representative sample of the Icelandic population. We document that the marginal propensity to consume (MPC) out of small windfalls, i.e., perfectly temporary unexpected income shocks, is larger than one for the average individual. Furthermore, we document that individuals who receive small windfalls increase their short-term unsecured consumer debt, such as overdrafts, in response. This borrowing response is prevalent for individuals having relatively little as well as a lot of liquidity, i.e., borrowing capacity. The larger-than-one MPCs are thus financed using expensive consumer debt that is then rolled over for a considerable period of time. For large windfalls we only observe small MPCs and no borrowing responses. We also document that individuals do not increase their savings in response to either small or large windfalls. Our findings point to overconsumption problems driving both high MPCs as well as large consumer debt holdings and are clean evidence against liquidity constraints as an explanation for high MPCs out of windfalls.

Selected presentations: FCA Household Finance Conference, Bonn Applied Micro Seminar, Copenhagen Workshop on New Consumption Data, AEA

Inactive working papers (read at your own risk):

Family Finances: Intra-Household Bargaining, Spending, and Financial Structure (joint with Arna Olafsson) click This paper aims to test recent influential theories proposing that differences in preferences of household members lead to agency problems reflected in overspending, indebtedness, and financial fee expenses at the household level. To do so, we use comprehensive transaction-level data from individuals within households. Observing individuals within households gives us a unique opportunity to empirically examine how individual revealed preferences over discretionary spending and individual patience affect spending and indebtedness at the household level. To deal with endogeneity, we use a fixed effects and instrumental variable approach, which helps us tackle both self-selection and common-shocks issues. We document that the share of household income received by the spender (impatient) spouse causally increases discretionary (total) spending at the household level, controlling for total household income. Moreover, we find that larger differences in household member patience increase debt and fee expenses at the household level. Our results are consistent with individuals having different preferences over spending and using expensive debt, which results in overspending and indebtedness at the household level.

Selected presentations:  Financial Research Association Conference in Las Vegas,  Families and the Macroeconomy Conference in Mannheim, AEA

Media coverage: Ideas at Work

The Consumption Response to Capital Gains: Evidence from Mutual Fund Liquidations (joint with Steffen Meyer and Alessandro Previtero) click Using a large sample of transaction-level data on all asset holdings, spending, and income from a German retail bank, this paper explores how individual consumption responds to realized capital gains. Our identification strategy exploits mutual fund closures, which are arguably exogenous to individual characteristics. We estimate the marginal propensity to consume (MPC) out of one dollar received from a forced sale event and find that it is approximately 30%. We explore how the MPC varies in age and income as well as over the business cycle and across interest rate regimes. We find a higher MPC for low-income investors, which appears consistent with standard life-cycle portfolio-choice models, though we do not find any differences in the MPC for young versus old investors. We also find that the MPC to be lower in recessions and decreasing in interest rates, which is surprising from a standard model perspective.

Selected presentations: Indiana University, the 3rd Annual CEPR Symposium, CSEF-IGIER Symposium on Economics and Institutions, EWFC at the WFA 2018, EMMMC, Workshop on New Consumption Data

Expectations-Based Reference-Dependent Consumption and Portfolio Choice: Evidence from the Lab (joint with Thomas Meissner, Philipp Pfeiffer, and Christopher Zeppenfeld) click In a lab experiment, we test standard consumption and portfolio choice predictions against those of expectations-based reference-dependent and hyperbolic-discounting preferences. The experiment consists of four periods. In the first period, subjects are endowed with experimental wealth. Then, subjects decide how much of their experimental wealth to “consume” by surfing the internet instead of performing an alternative monotone task. To consume in future periods, they either store their wealth safely or invest it into a risky lottery. The main predictions of reference-dependent preferences, which stand in contrast to those of standard and hyperbolic-discounting preferences are: First, the consumption share is decreasing in the investment outcome. Intuitively, the agent delays painful cuts in consumption to let his expectations-based reference point decrease. Second, the portfolio share is decreasing in the outcome. The agent increases his risk exposure in bad states to not realize too many loss feelings about future consumption. Third, the agent’s behavior is not time consistent. The agent likes to increase his consumption and risky asset holdings above expectations today, but considers his expectations when making plans about tomorrow.