Abstract: When a public entitlement is grandfathered only while an asset is held, selling forfeits what holding preserves. I study Beijing’s January 2019 title-vintage rule, under which transfers moved buyers from a single-school mapping to multi-school assignment. In 54 exactly linked communities, formerly key and ordinary communities both begin near three platform completions per 1,000 Lianjia-recorded housing units. Over the next sixteen weeks, ordinary communities rise to about six while formerly key communities remain near three. The raw exposed-relative PPML ratio is 0.57–0.61; seasonal benchmarks imply 0.42–0.67, with the largest gap against an unusually strong 2018. The negative sign survives frozen samples, homogeneous completion clocks, count estimators, and every school and community deletion. Public active-stock and asking-price comparisons move less, conditional on unverified crawler comparability, but do not identify a listing hazard. A search-and-bargaining model shows how a sale-triggered entitlement reset can remove marginal matches without requiring a commensurate selected-price response.

Abstract: Variance compresses favorable and adverse earnings realizations into one scale measure. I use public CPS earnings histories and strictly later SIPP portfolios to separate lower-tail distance, upper-tail distance, total span, and orientation. Among 27,752 annual transitions beginning with no reported stock-and-mutual-fund holdings, a one-standard-deviation wider lower tail predicts a 2.71-percentage-point lower probability of positive holdings one year later and a wider upper tail a 1.92-point higher probability; the pair remains jointly distinguishable from zero under a 200-draw CPS first-stage sensitivity. With 99,999 cell-level wild-cluster draws, the lower and upper coefficients have two-component adjusted q-values of 0.019 and 0.087, and the signed composition contrast has p = 0.033. An exact rotation leaves total span imprecise and composition negative. Recomputing the CPS tails after excluding destination records classified as self-employed sharply attenuates the lower estimate. Across alternative zero-endpoint constructions, composition retains its direction and conditional precision; the standalone upper estimate remains positive but is imprecise in one construction. Historical group distributions are assigned rather than elicited, so the estimates are descriptive.

Abstract: Using transaction-level data on 189,530 Chinese retail investors over the 2013–2016 boom and bust, we measure experience as counts of large realized gains and losses and find that it reshapes the disposition effect, the tendency to sell winners and hold losers, in an asymmetric way. A two-standard-deviation increase in accumulated large losses raises the propensity to sell winners rather than losers by about 23% of the baseline effect, while an equivalent increase in gains lowers it by about 21%. Composition-matched simulations show the asymmetry is not a mechanical artifact of building experience from past trades. To explain both the baseline effect and this asymmetry, we build a memory-based recall model in which similarity-weighted retrieval of past outcomes guides selling. A single retrieval channel reproduces all three patterns without shifting preferences.

Abstract: Repeated elicitation is commonly used to improve the measurement of economic preferences. I test that premise using 56,050 intertemporal-choice responses from 6,808 Japanese households. Responses observed through 2015 predict financial wealth in 2022–2024 conditional on 2016 wealth, allowing a common-sample comparison of one response, averages, response distributions, and split-block IV. Averaging produces a negative displayed-rate gradient, and split-block IV makes it steeper. Yet the gradient is driven by the questionnaire’s 40% endpoint: capping that category makes the slope imprecise, and flexible category models reject displayed-rate linearity. Moreover, post-first responses add virtually no out-of-sample prediction beyond the first category. Repetition therefore makes persistence and coding assumptions visible, but does not produce a demonstrably better predictor of future wealth. The estimates are predictive rather than causal. The result is a boundary condition for treating repeated survey responses as increasingly precise measures of latent preferences.

Abstract: An involuntary job loss can reduce household liquidity, while any job ending can alter how employer-plan balances are handled. I use two regimes of the Survey of Income and Program Participation and direct reports of pension or retirement lump sums and rollover plans to study annual non-rollover distributions. A conventional permanent-layoff–no-layoff comparison yields a 1.74-percentage-point association, relative to a 0.79-percent mean. Holding the sample and specification fixed produces a 1.55-point permanent-layoff contrast and a 0.40-point involuntary–voluntary separation contrast. Among prior workers reporting a 401(k) or thrift account, both voluntary and involuntary separators have higher non-rollover incidence than continuers, while their difference is imprecise. Outcome accounting shows that receipt rises for both groups and that most of each receipt difference is accounted for by non-rollover. Non-rollover is therefore elevated around separation broadly, and the treatment-and-comparison definition is central to interpreting a layoff coefficient.

Abstract: Did industrialization under central planning leave a distinctive marriage-age profile? We study China’s Third Front with the 1982, 1990, and 2000 population censuses. In 1982, a one-standard-deviation increase in the heavy-industry share is associated with a 4.23-percentage-point larger gap in never marriage between ages 20–24 and 30–34 inside Third Front provinces than elsewhere; predetermined siting controls reduce the differential to 2.24 points. Retrospective first-marriage histories observed in 2000 yield similar pseudo-1982 point estimates under birth-year and completed-age indexing. The point-estimate alignment survives on identical respondents, although uncertainty permits meaningful differences and current residence cannot recover the 1982 population. Campaign-aligned comparisons show an early marriage deficit and subsequent catch-up within industrial Third Front areas, but no robust additional shift relative to other areas. Education shares account for 15.5 percent of a comparable young–old gap. The paper establishes a persistent resident-population marriage-age gradient and a transparent bridge from census stocks to cohort histories, while showing why neither identifies the campaign’s causal effect.

Abstract: How do labor-market conditions at school leaving shape what households later save for? I link 4,120 Japanese women observed in 29 waves of the JPSC to the national job-openings ratio at predicted graduation. Moving from late-bubble to trough entry conditions predicts 4,750 yen more monthly saving among married women with complete saving reports, about 7 percent of the sample mean. Roughly 87 percent of this response is designated for family-common, child, or other purposes rather than either spouse. Adverse entry also predicts fewer durable purchases and more discouraged borrowing, consistent with a defensive financial posture. Among positive savers, spouse-designated shares fall, but the wife–husband split does not move precisely; the share pattern is therefore an amount-decomposition result, not evidence of a transfer between spouses. The allocation contrast is robust to cohort clustering and cohort-score HAC inference and is extreme among circular treatment rotations, although studentized rotations and within-year designs are less decisive. The results extend labor-market scarring from household balance sheets to the purposes attached to saving.

Preliminary abstract: Does the same investor treat a loss differently when it is hers rather than an algorithm’s? We study 100,000 investors on Alipay’s Ant Fortune platform (September 2020 to September 2024) who simultaneously hold a self-directed mutual-fund portfolio and a robo-advisory account (BangNiTou). Comparing the same investor across two accounts that differ only in who chose the holdings, we find that the response to losses does not weaken under delegation—it reverses sign. On her own account she holds and adds to losers (averaging down); on the delegated account she sells losers and chases winners. A single loss-aversion parameter cannot produce this sign flip. It follows, we argue, from delegation changing what a return means: on her own account a loss is processed through realization utility, while on the delegated account it is both a signal about the algorithm and a source of dissonance she resolves by exiting. We estimate the model structurally and quantify the return investors forgo through self-directed loss aversion. Full empirical results remain in progress.