Where there are price controls, there is hoarding. Families in Venezuela stockpiled goods during years of controls and shortages, and similar responses followed price controls in Argentina and shortages during the COVID-19 pandemic. Yet experimental evidence on how price distortions interact with storage has been scarce.
Castro, Martinelli and Brown model hoarding as a rational response to the risk of rationing. Storage can arise for two different reasons: defensive storage insures buyers against expected future rationing, while rent-seeking storage lets them buy at a regulated price in anticipation of consuming or reselling after prices are freed. Both lead to hoarding, but with different consequences for efficiency.
Under a price cap that participants knew would be lifted, the shortage rate rose from 12% in round 1 to 34% in round 10, and the price more than doubled, from 50 to 109, once the cap ended.
Three markets in the laboratory
The experiment brought together 288 students at George Mason University in 22 sessions. In each round, groups of eight participants acted as buyers in a two-period market with nine automated sellers programmed to follow equilibrium strategies. Production costs quadrupled in the second period, so buyers had a reason to buy early: in period 1 they could purchase a second unit and store it for period 2 at a cost of 10.
Sessions were assigned to one of three environments. In the Free Market, prices were unrestricted in both periods. Under the Anticipated Price Cap, the first period was free and a cap of 50 applied in the second, so buyers expected to be rationed later. Under Liberalization, a cap of 50 bound in the first period and was removed in the second. Using automated sellers isolates the buyers' side: departures from the predictions can be attributed to buyer behavior.
Liberalization: growing shortages, then a price jump
When a binding cap was expected to be lifted, shortages in the capped period worsened with experience. The shortage rate rose from 12% in round 1 to 34% in round 10, and it increased in every one of the seven liberalization sessions. Participants appear to learn to position their bids to obtain units at the controlled price before liberalization.
Once the cap was removed, the average price rose from 50 to 109, more than double, in line with the model's prediction. Buyers who managed to store paid 60 including the storage cost for goods that cost about 109 in the free second-period market.
An anticipated cap: too little storage
Theory predicts that buyers who expect to be rationed later should bid more and store more today: 3 units per market at a first-period price of 88 to 90. Participants stored 2.31 units and paid 73 on average, almost exactly what they did in the free market (2.36 units at 73.67). The additional rationing incentive had no detectable effect on storage, as if participants underestimated how severe future rationing would be.
The shortfall is not a general reluctance to store, and measured risk preferences do not account for it. The paper discusses two candidate explanations, limited strategic reasoning about how others' storage affects future rationing and preferences over when uncertainty is resolved, but the experiment was not designed to tell them apart.
The cost of price controls
Total surplus, the value of consumption minus production and storage costs, was highest in the Free Market (701 per market), about 8% lower under the Anticipated Price Cap (647) and about 13% lower under Liberalization (609). Liberalization was the lowest-surplus environment because it combined current rationing, which randomly excludes some high-value buyers, with rent-seeking storage and a sharp price increase after the cap was lifted. These losses did not shrink with experience.
The design has limits. Automated sellers rule out supply-side responses such as quality degradation or withheld supply, participants were university students, and the experiment was not preregistered. The authors suggest varying storage costs, eliciting beliefs about rationing directly and introducing strategic sellers as next steps.
Findings
- Shortages only where caps bind Shortage rates were near zero in every period without a cap, 25% in the capped first period under liberalization and 60% in the capped second period under an anticipated cap.
- Liberalization worsens shortages Under a cap expected to be lifted, the shortage rate rose from 12% in round 1 to 34% in round 10, and the price rose from 50 to 109 after liberalization.
- Too little defensive storage Facing a future cap, buyers stored 2.31 units per market instead of the predicted 3 and paid 73 instead of 88 to 90, much as in the free market.
- Price controls cost surplus Total surplus was 701 in the free market, 647 under an anticipated cap and 609 under liberalization, and the losses persisted with experience.
From evidence to action
Look beyond today's shortage
The consequences of a price distortion depend not only on the transactions it prevents today but also on the forward-looking behavior it induces, including storage in anticipation of tomorrow's market.
Expected liberalization has its own costs
When a cap is expected to be lifted, buyers increasingly compete for goods at the controlled price, and in the experiment this environment produced the lowest total surplus.
Test why anticipation falls short
Future experiments can vary storage costs, elicit beliefs about rationing and add strategic sellers to explain why buyers under-respond to anticipated rationing.
Abstract
We study how current and anticipated price distortions affect storage decisions in a two-period model. Both anticipated rationing and anticipated price liberalization create incentives to hoard, though for different reasons. We test these predictions in the laboratory using three treatments: a free market with rising costs, an anticipated price cap, and liberalization following a price cap. The model successfully predicts the broad patterns of shortages and prices across treatments. Under liberalization, shortages become increasingly severe as participants gain experience with the regulated market, and prices more than double once the cap is lifted. When a price cap is anticipated, however, participants store less than predicted and bid too little for storage, as if they underestimate the severity of future rationing. Both price-control regimes entail substantial welfare losses with respect to the free market, with liberalization producing the lowest surplus as rent-seeking rather than defensive storage becomes the dominant motive for hoarding.
Castro, Edgar, César Martinelli, and Nicholas Brown. 2026. "Price distortions and hoarding: An experiment." BEER Lab Working Paper BEER-WP-2026-02, Tecnológico de Monterrey.
Research team
Edgar Castro, César Martinelli and Nicholas Brown. Edgar Castro is at Tecnológico de Monterrey; César Martinelli and Nicholas Brown are at George Mason University. The experiment was run with the participant pool of the Interdisciplinary Center for Economic Science (ICES) at George Mason University. An earlier version of this paper was Chapter 1 of Castro's doctoral dissertation at George Mason University.