Research
Research
Since the 1980s, the United States has experienced a pronounced saving glut of the rich, a large accumulation of assets among the top 1% of earners. We argue that this development partly reflects a shift in the financing of redistributive policies, from inflationary finance in the 1960s and 1970s to debt finance backed by future taxation beginning in the early 1980s. The central insight is that, in the presence of a progressive tax system, a switch from inflationary to debt financing leads to debt accumulation by top earners. We develop a New Keynesian model with borrowers and savers in which redistributive transfers can be either funded or unfunded. Unfunded transfers generate fiscal inflation that erodes the real value of public and private debt, redistributing wealth through asset revaluation effects. Funded transfers, by contrast, are financed through future taxes borne primarily by high-income households, which respond by accumulating claims on both households and the government. Using a structurally estimated version of the model, we find that the shift from unfunded to funded redistribution in 1980s contributed significantly to the subsequent saving glut of the rich.
How do interest rates affect durable-goods markets in general equilibrium? I develop a directed-search model of differentiated durables in which state-dependent failure rates sort ex ante identical buyers and sellers across submarkets. Buyers' optimal policy partitions the state space into action and inaction regions with target quality monotonic in the holding, and market tightness is decreasing in quality. Calibrated to U.S. vehicle microdata, the model predicts that a decline in the interest rate triggers a short-run overshoot of durable expenditure; firms reallocate production toward higher-quality submarkets, and the economy converges to a stationary equilibrium of less frequent, higher-quality purchases.
We study optimal taxation in economies with general equilibrium market clearing, where agents with privately known labor skills and entrepreneurial abilities choose between deterministic labor income and risky firm operation. The government observes labor income and realized dividends but not effort or technology shocks. We formulate the multidimensional screening problem as a lottery-based linear optimization, accounting for global incentive constraints, fixed costs and other non-convexities. Optimal policies exhibit tax breaks, which can render net taxes negative, for agents with intermediate entrepreneurial abilities and labor skills above a threshold. General equilibrium strengthens this effect under decreasing returns, as labor-market clearing requires sufficient entry into entrepreneurship, further increasing subsidies for agents with high worker options. In a calibrated U.S. economy, optimal taxes are lower and can be negative for low-profit realizations. Subsidies rise when risk declines and when the frequency of high-ability entrepreneurs in the population diminishes. Global incentive constraints bind only near the occupational frontier, amplifying tax breaks for marginal entrepreneurs.
To study how peer-dependent replacement interacts with embodied innovation, I introduce a vintage-capital economy in which capital users choose when to replace old vintages and capital-goods producers choose how fast to improve the frontier. Users interact through a product-market quality benchmark built from the installed capital of their peers; producers interact with users through the market size created by replacement demand. The central mechanism is a two-way feedback between replacement and innovation: peer upgrading raises the private cost of operating old capital, while more replacement thickens the market that rewards frontier improvement. When the regular equilibrium correspondence is monotone on the admissible interval, the balanced-growth path is unique. When peer feedback creates a regular turning point, the same primitives support multiple replacement-growth regimes through a saddle-node mechanism, with slow replacement and weak innovation or fast replacement and rapid innovation. In the calibrated economy, belief shifts across continuation branches generate synchronized replacement waves.
Bank markups have risen substantially, dispersion across banks has increased, and large banks now charge higher markups than smaller institutions. We develop a general equilibrium model in which persistent borrower and depositor relationships endogenously generate heterogeneous bank market power. Banks are dynamic two-sided intermediaries that compete for customers while inheriting partially captive borrower and depositor bases. Relationship capital creates a trade-off between current margins and future franchise value, generating endogenous loan markups and deposit mark- downs that vary across banks. Regulatory constraints and costly external equity make market power on one side of the balance sheet affect pricing on the other, linking deposit markdowns and loan markups. Quantitatively, two-sided bank market power has sizable macroeconomic implications, reducing financial intermediation and lower- ing aggregate output. Policy-rate changes alter franchise values, leading banks with different customer bases to adjust loan and deposit rates differently.
A frontier technology is useful only as part of a system: the return to implementing it in one industry depends on implementation in others. I develop a dynamic multi-sector model in which producers license the frontier from innovators, implementation choices are strategic complements over a directed network, and license fees finance the research that creates the next frontier. When network feedback is strong, pessimism is self-fulfilling and compounds: weak implementation lowers the innovation prize, research stops, and technology gaps erode until coordination becomes impossible. Technology gaps determine whether coordination is possible; hype only selects when it is. Measuring the network from the technology content of U.S. job postings and identifying implementation responses with granular firm-level instruments, I find that the estimated economy admits both a collapsed and a coordinated branch. The coordination failure costs 0.4 percentage points of annual TFP growthand implementation support dominates equal-budget R&D subsidies.