A Case for Modernizing Fiscal Policy
John Maynard Keynes's General Theory of Employment, Interest, and Money has shaped economic and governmental thought for nearly a century. Its central insight, that government spending can jolt a stalled economy back to life through a multiplier effect, remains sound. But the world in which that insight was applied no longer exists, and two of its load-bearing assumptions have quietly failed. The first is economic: the multiplier itself has weakened. The second is political: the officials Keynes trusted to wield fiscal policy responsibly have not lived up to that trust. Fixing fiscal policy for the twenty-first century means addressing both failures directly, and doing so requires a new institution, a Federal Stimulus Reserve, built with the independence of the Federal Reserve but a mandate for fiscal, not monetary, policy.
Start with the economics. Keynes's multiplier worked because in the 1930s and 1940s, manufacturing and farming accounted for over sixty percent of U.S. employment, and a dollar spent on goods circulated back to American workers almost automatically. Today, those two sectors account for under twenty percent of employment; services dominate, and services spending is exactly what consumers cut first in a downturn. Meanwhile, global supply chains mean that much of what stimulus dollars buy is manufactured abroad, so spending that once supported American factory workers now supports workers in other countries. The result is a fiscal dollar that does measurably less domestic work than it did in Keynes's time, not because the theory is wrong, but because the economy it was built for has changed underneath it.
Now the politics, which is the older and more stubborn problem. Keynes wrote from inside a British civil service and political class that he had good reason to trust: well-educated, empire-tested administrators managing a quarter of the world's landmass with real competence. Even so, contemporaries like Joseph Schumpeter doubted that elected officials would ever reliably choose the nation's long-run interest over their own short-run survival. That skepticism has aged better than Keynes's optimism. His framework depended on politicians doing two things: authorizing deficit spending during a crisis, and then imposing fiscal discipline once the crisis passed. American politics has shown a durable talent for the first and a durable aversion to the second. The 2009 Recovery Act is the clearest recent case, too small, too slow to disburse, shaped as much by political horse-trading as by economic need, and no meaningful fiscal correction followed once the economy stabilized. Austerity is unpopular, and elected officials facing reelection have never had much appetite for unpopular.
The United States has already solved a version of this problem once. Monetary policy used to be just as exposed to short-term political pressure as fiscal policy is today, which is precisely why the Federal Reserve was built with independence from Congress and the White House. That independence is now credited as a central pillar of postwar U.S. price stability. Fiscal policy has no equivalent. It is still authored entirely by officials whose incentives point toward action during a crisis and inaction afterward. The comparison is not exact, fiscal policy involves spending choices that carry legitimate democratic stakes in a way that setting an interest rate does not, but the underlying logic transfers: some degree of insulation from short-term political incentives would make fiscal policy more effective, not less democratic, in the ways that matter most during a crisis.
The proposal that follows from both problems is a Federal Stimulus Reserve: an institution modeled structurally on the Federal Reserve System, independent, Washington-based, with regional branches, but carrying a fiscal rather than monetary mandate. Where the Fed manages the money supply, the Reserve would manage the deployment of stimulus capital: fast, insulated from political pressure, and built around continuous economic research rather than periodic congressional negotiation. Its research function, modeled loosely on the St. Louis Fed's role within the Federal Reserve System, would track the economy in real time and study which forms of stimulus generate the largest domestic multiplier, addressing the first problem, the weakened multiplier, directly. Its independence from Congress would address the second problem, political unreliability, by giving the country an institution that can move at the speed a crisis actually requires, rather than the speed Congress is willing to move.
An institution is only as credible as its funding, and this is where the proposal needs to be most disciplined. The obvious funding source is the Federal Reserve's own open market operations profits, which currently flow to Treasury and could instead be redirected to the Reserve. But that source alone is not dependable enough to build an institution on. Fed profits come from the spread between what it earns on its bond holdings and what it pays out on reserves, and that spread compresses, sometimes inverts, during periods of rising interest rates. In 2022 and 2023, the Fed ran an operating loss and remitted nothing to Treasury at all. Rate-hiking cycles often precede or coincide with downturns, which means a reserve funded solely by Fed profits would be most likely to run dry in exactly the years before it is needed most. A credible funding structure has to account for that.
The more durable design is a dual-source structure: a standing annual appropriation, sized as a small fixed percentage of GDP, as the primary and predictable funding stream, similar in spirit to how the FDIC's Deposit Insurance Fund is capitalized through steady industry assessments rather than favorable market timing, with Fed open market profits layered on top as a secondary, opportunistic source whenever they are available. This guarantees the Reserve accumulates capital on a predictable schedule regardless of the interest rate environment, while still capturing Fed profits as a bonus in the years they materialize. It is also worth being explicit about why a pre-funded reserve is worth this complexity at all, rather than relying on existing automatic stabilizers like extended unemployment insurance, which already trigger by formula. The answer is speed: automatic stabilizers are contingent authorizations that still require fresh appropriation once triggered, while a pre-funded reserve holds capital that already exists and can be released in days rather than months, the same operational advantage that lets the Fed act on monetary policy without waiting on Congress.
Funding solves the speed problem; it does not solve the leakage problem. A well-capitalized Reserve that simply mails out checks still runs into the same weakened multiplier described above, dollars spent on imported goods or discretionary services that contract during downturns. Addressing that requires a more targeted delivery mechanism than a traditional stimulus check: a direct-payments system with built-in spending conditions, where eligible recipients receive funds that can only be spent with certified vendors on certified goods and services. Certification would prioritize U.S.-manufactured goods, goods with meaningful domestic supply chain content, and service providers meeting domestic employment criteria, essentially applying the existing EBT certification model to a broader, recession-triggered payment system.
The technology for this already exists. Real-time transaction categorization, the same infrastructure that already powers credit card cash-back categories, can verify at the point of sale whether a purchase qualifies. Delivery could run through a mobile app or an NFC-enabled card, with a map function directing recipients to nearby certified vendors, which doubles as a way for certified businesses to advertise their eligibility during a downturn and grow the certified network faster. A two-round structure could further sharpen the incentive to spend rather than save: an initial unrestricted payment, followed by a larger second payment contingent on demonstrated spending of the first through certified channels. None of this changes how much money enters the economy. It changes where that money goes once it arrives, which is the entire point, recovering as much of the lost multiplier as a targeting mechanism reasonably can.
Even a Reserve that solves both the speed problem and the targeting problem still leaves the oldest problem in place: nothing compels responsible fiscal policy once the crisis has passed. That failure is not a character flaw in individual politicians; it is the predictable output of an incentive structure that rewards visible action during a crisis and punishes visible austerity afterward. Legislating that incentive away is not realistic. Designing around it is. The same institutional structure that lets the Reserve deploy capital quickly can also generate the political cover elected officials need to act responsibly once the crisis ends.
The mechanism is symmetrical with how the Reserve deploys funds in the first place. When it releases stimulus capital, it simultaneously sends Congress a prewritten bill, populated with the specific dollar figures its research indicates are necessary, for an up-or-down vote with no room for unrelated riders. When the crisis passes, the Reserve sends a second prewritten bill specifying the intermediate-term budget correction, built the same way, on the same evidentiary basis. In both cases, Congress's role narrows from authoring fiscal policy to ratifying expert-generated numbers, which shifts public pressure toward action rather than inaction. This is the same professionalization the Federal Reserve achieved for monetary policy, applied for the first time to the fiscal side of the ledger, and it is the piece of the proposal that most directly answers the oldest criticism of Keynes: that deficit spending is rarely followed by responsible policy afterward.
The institutional logic developed above, that crisis response works best when the government has in-house, professionalized capacity to act quickly rather than assembling it under pressure, applies beyond the Reserve itself. It applies just as clearly to a gap that already exists inside the Federal Reserve. The Federal Reserve Act of 1913 was designed for the financial system of 1913, and during COVID-19 the Fed found itself needing capabilities that act had never anticipated: buying newly issued corporate bonds, existing investment-grade debt, and credit-market ETFs to keep credit flowing. Lacking the in-house capacity to execute those purchases, the Fed contracted the work out to BlackRock, handing a single private asset manager an extraordinary degree of influence over the credit markets at the exact moment that influence mattered most.
That outsourcing was a symptom of the same underlying gap the Reserve is designed to close: the government builds emergency capacity only after the emergency has already arrived. The fix follows the same pattern. A dedicated Asset Management Division within the Federal Reserve Bank of New York, built to execute and manage these expanded asset classes directly, would let the Fed respond to the next crisis with its own capacity rather than a contractor's. Paired with a longer-term investment in system-wide asset monitoring, something approaching the ambition of BlackRock's own Aladdin platform, built for public rather than private use, this would give the Fed the same kind of institutional readiness for credit-market intervention that the Reserve is designed to provide for fiscal stimulus.
A Reserve designed to make deficit spending faster and better targeted only strengthens the case for taking the resulting debt seriously, and this is the area where American political debate is least serious. Conservatives tend to call for repayment through immediate, severe cuts; liberals tend to avoid the subject almost entirely. Neither produces a workable plan, and the absence of any credible plan is itself costly, the United States' 2011 credit downgrade reflected doubt about political will, not financial capacity, since the U.S. is one of only four countries, alongside France, Canada, and Australia, to never miss a debt payment.
A more workable framework treats the debt the way a long-lived institution should: not as a ten-year obligation to be paid off at painful speed, but as a fifty-to-one-hundred-year commitment, sized the way a thirty-year mortgage is sized relative to a ten-year one, large enough to matter, small enough per year to leave room for the education and infrastructure spending that responsible governance still requires. The United States has a time horizon no individual borrower has; a repayment plan should be proportioned to that horizon rather than compressed to fit an election cycle. The credibility of adopting any such framework, even a slow one, is what restores confidence, not the speed of repayment itself.
Each piece of this proposal answers a specific failure in how the United States currently manages economic crises: an institution with Fed-like independence to solve the speed problem, a dual-source funding structure to make that independence financially credible, a targeted payments system to recover the multiplier that globalization eroded, an automated legislative mechanism to enforce the fiscal discipline politics has never reliably supplied, an in-house asset management capacity to close the same readiness gap inside the Fed itself, and a long-run debt framework proportioned to the country's actual time horizon rather than its electoral one. None of these ideas is a repudiation of Keynes. Each is an attempt to update the machinery around his central insight for an economy, and a political system, that have both changed substantially since 1936. Keynes set out to tame the business cycle; taming it entirely is probably not possible, but building institutions capable of responding to it faster, more precisely, and more responsibly than the current system does remains well within reach.