Diego Escobari

This paper finds empirical support to systematic peak-load pricing in airlines —higher fares in ex-ante known congested periods. It estimates a congestion premia and supports the main empirical prediction in Gale and Holmes [Gale, I., Holmes, T., 1993. Advance-purchase discounts and monopoly allocation of capacity. American Economic Review 83, 135–146] —less discount seats on peak flights.

We show how a monopoly airline uses two ticket types, refundable and non-refundable, to screen consumers with different willingness to pay. Our theoretical model suggests that the difference between these two fares consists of refundability value and price discrimination, and the fare difference diminishes as risk-averse passengers learn about their individual demand uncertainty. Using an original dataset, we find that, after controlling for unobserved seat- and flight-specific characteristics, the empirical evidence from the U.S. airline industry supports our theory.

Airlines sell tickets in advance when there is uncertainty about the aggregate demand. Dynamic pricing is important because in low demand flights unsold tickets are of little value after departure and in high demand flights some high valuation consumers may be rationed. Using a unique panel of U.S. airline fares and seat inventories at the ticket level, this paper shows that airlines learn about the aggregate demand as sales progress and departure date nears. The paper finds strong empirical support to stochastic peak-load pricing, showing that carriers set higher fares when demand is more likely to exceed capacity.

This paper tests the empirical importance of the price dispersion predictions of the Prescott-Eden-Dana (PED) models. Equilibrium price dispersion is derived in a setting with costly capacity and demand uncertainty where different fares can be explained by the different selling probabilities. The PED models predict that a lower selling probability leads to a higher price. Moreover, this effect is larger in more competitive markets. Despite its applications to several important market phenomena, there exists little empirical evidence supporting the PED models, mostly because of the difficulty of coming up with an appropriate measure of the selling probabilities. Using a unique panel of U.S. airline fares and seat inventories, we find evidence that strongly supports both predictions of the models. After controlling for the effect of aggregate demand uncertainty on fares, we also obtain evidence of second degree price discrimination in the form of advance-purchase discounts.

Abstract available upon request.

Despite the empirical evidence that shows that tax audits are imperfect, the standard assumption in the literature is that all tax evasion is detected once the taxpayer is audited. This article models the existence of imperfect detection of tax evasion motivated in a corrupt tax administration. The model finds that corruption gives auditors incentives to discover more evasion. Consistent with previous literature, fines and audit probabilities both have a positive effect on compliance. However, failing to consider the effect of imperfect detection will overestimate this positive effect. Giving inspectors a share of evasion detected will increase compliance levels, while increasing their lumpsum income will reduce it. Higher compliance can as well be achieved by having more efficient inspectors or lower corruption levels.

This paper tests for convergence using time series data from nineteen Latin American countries over the period 1945-2000 with two existing notions of convergence and with a herein proposed testable definition of β-convergence. The results show that only one pair of countries, Dominican Republic and Paraguay, were found to pair wise converge according to Bernard and Durlauf (1995) definition. Additionally, more evidence of stochastic convergence was found when allowing for structural breaks using the two-break minimum Lagrange Multiplier unit root test proposed by Lee and Strazicich (2003) than when no breaks were considered. Stochastic convergence within groups showed that there is more evidence of convergence within Central America and Caribbean economies than within South American economies. Only one country, Dominican Republic, was found to comply with the neoclassical growth models' conditions for income convergence, implying that the shocks to its real output per capita are temporary and that as relatively poor country is catching up with richer economies.

Abstract available upon request.

Abstract available upon request.