This is rooted in the findings in Kageyama
（2012） , which empirically showed that the relationships between LEGAP and these happiness indicators are bidirectional. In one direction, LEGAP negatively affects both HPN and HPGAP . An increase in LEGAP raises women ’ s widowhood ratio, and, since widows are, on average, less happy, it lowers women ’ s average happiness, HPN, and HPGAP . We call this effect the “ marital-status composition effect ” as the marital-status composition plays a central role.
This has two implications. First, parents use the same income-dependent relative preferences to measure their children’s well-being. Second, since parents obtain their own utility from consumption in adulthood, they care about their children’s consumption in their adulthood. To capture this latter aspect, I assume that parents are concerned with children’s growth that signals the children’s consumption in adulthood and measure children’s growth with educational output.
We next turn to the price elasticities implied by the demand estimates— the percentage change in the probability of choosing the audit firm result- ing from a 1% increase in audit fees. Here, we follow standard practice in discrete choice demand estimation by imposing that changes in price have the same impact on utility for all products (i.e., a common α). The eco- nomic logic of this standard assumption is straightforward: paying a given amount more in the form of higher prices has the same effect on the con- sumer’s utility regardless of what choice that expenditure was put toward (this follows naturally from the notion that the opportunity cost of those expended dollars to the client firm is the same regardless of which auditor received those dollars). Note, however, that imposing a common price coef- ficient across all choices does not impose that price elasticities are the same alternative specification do equally well in predicting the choices of clients for the Big 4 audit firms, implying that the main specification well explains preferences for the Big 4 through- out the client size distribution. To further evaluate the robustness of the estimates, we replace nonmatch prices with “what if” prices based on three years of tenure to capture any effects of low-balling. The parameter estimates for this alternative specification are similar to the main results. In addition, we replaced audit fees with total fees and again found similar results.
To perform our counterfactuals we need some identiication assumptions. The empirical model identiied only the ratio F /(c B − c A ). Based on our external estimate, we assume that F = €2,300, implying that c B − c A is about €0.60. The results were similar for higher values of ixed costs with correspondingly higher variable cost savings. The empirical model also does not identify the travel cost per unit of distance k (in €/km) from the price parameter a. In our welfare analysis, a high assumed value of k implies a high weight to consumer surplus relative to producer surplus. We therefore use two alternative values k = €0.10 and k = €0.25. The higher value is commonly used by companies and tax authorities to reimburse travel costs. The lower value roughly cor- responds to Gowrisankaran and Krainer’s (2007) estimate of ATM travel costs (using a different model and data ). We focus our discussion on the results for k = €0.25 and present the results for k = €0.10 as a robustness check in Table A3 of Web Appendix B.
θ i and weakly decreasing in q. Horta¸csu and Kastl (2012) provide a formal method to test
for the null hypothesis of private values in the Bank of Canada’s three-month treasury-bill auctions, and do not reject private values in that application. Their test method relies on the specific institutional setup in the Canadian treasury market, which is not present here. We do however argue that their results provide support for our assumption of independent private values in the context of government debt auctions. It can reasonably be argued that banks have idiosyncratic shocks to their liquidity needs due to deposit flows and the corresponding reserve requirements. The assumption we impose in our empirical work is that these shocks are independent conditional on observed macro and secondary market conditions. In the sequel we drop the index i, as there is no danger of confusion
The irst instruments we use are based on the net amount of charged-off loans by a bank. Loan charge-offs measure the net value of loans and leases that were removed from the bank’s balance sheet because of uncollectibility, and are one mea- sure of the performance of a bank’s loan portfolio. We include bank ixed effects in the speciication, so our results are not driven by the fact that banks which give bad loans also offer poor services that make them unattractive to depositors. Instead, our instrument is identiied from changes in loan charge-offs within a bank over time. We use two types of charge-offs: those for all loans, and charge-offs for real estate secured loans. Quarter ixed effects absorb any aggregate activity, which would affect loan performance, and would also change depositors’ preferences. Moreover, loans that are written off in a given period have been made in the past, so we exploit the variation in the quality of the loans the bank has made in the past to generate variation in inancial distress in the present. Therefore, the possibility is small that changes in loan charge-offs within a bank over time would measure the services that depositors obtain from a bank.
k 兲 g共c j 兲
1 ⫺ 1 ⫺ . . . ⫺ j ⫺1
. Note that search cost distribution densities g(c), evaluated at the cutoff values for funds offering lower utility than j (i.e., k ⬍ j), affect fund j’s demand elasticity. To see why, consider investors’ reactions to an increase in the price of fund j. The price hike decreases u j . This has two distinct effects on the critical search
percentage of commercial and industrial (C&I) loans relative to total loans (RCFD2122). % Consumer Loans is the percentage of consumer loans relative to total loans (RCFD2122). 30-89PD Ratio is the ratio of total loans that are 30-89 days past due (RCFD1406) to total loans. NPL Ratio is the ratio of nonperforming loans (nonaccrual) (RCFD1403) and 90 days or more past due (RCFD1407) to total loans. OREO Ratio is other real estate owned (RCFD2150) divided by total assets. Unused Commitment Ratio is unused commitments divided by unused commitments and total loans. Tier 1 Capital Ratio is the ratio between Tier 1 (core) capital and total risk-weighted assets. Leverage Ratio is the ratio between Tier 1 (core) capital and (adjusted) total assets. % Core Deposits is total core deposits (transaction accounts + savings deposits + time deposits less than $100,000) divided by total deposits. Distance is the average pairwise distance (in 100-mile increments) between all pairs of branches of the failed bank and potential acquirer. State Bank is an indicator variable that takes the value of one if the bank is regulated by a state regulator. Multibank BHC is an indicator variable that takes the value of one if the bank is part of a bank holding company that owns more that one commercial or savings bank. Distance (% Res. Loans) is the absolute difference between the failed bank’s and the potential acquirer’s percentage of total loans held in residential loans. Distance (% CRE Loans) is the absolute difference between the failed bank’s and the potential acquirer’s percentage of total loans held in CRE loans. Distance (% CI Loans) is the absolute difference between the failed bank’s and the potential acquirer’s percentage of total loans held in C&I loans. Distance (% Cons. Loans) is the absolute difference between the failed bank’s and the potential acquirer’s percentage of total loans held in consumer loans. HHI is the average increase in local deposit market concentration that would result from potential acquirer j acquiring the branch network of failed bank i. HHI ranges from 0 to 1,000, where 0 indicates a merger that does not increase local market concentration and 1,000 indicates a merger that transforms a perfectly competitive local market into a local monopoly.
al. (2014) also show that optimal bid shading in these auctions also distorts the efficiency of the allocations, and thus a general ranking of expected revenues from discriminatory and uniform price auctions can not be made without knowledge about the specific features of bidder demand.
Given the theoretical vacuum, a variety of empirical approaches have been employed to as- sess the efficacy of Treasury auction mechanisms. The Treasury’s own study of this question, as reported by Malvey and Archibald (1998), was based on experimentation with the uniform price format for 2- and 5-year notes. To assess the revenue properties of the uniform vs. the status-quo discriminatory format, Malvey and Archibald calculated the auction-when-issued rate spread, and did not statistically reject a mean difference across the different auction formats. However, they note that the uniform price auctions “produce a broader distribution of auction awards” across bidders, and especially a lowered concentration of awards to top primary dealers.