Yves here. Trump has been particularly open and aggressive in his regular whinges about how the Fed had been undermining him by refusing to cut interest rates. We’ve repeatedly pointed out that, contrary to the deeply-held belief of many politicians, putting money on sale does not do much to stimulate the economy. Businessmen expand their operations based on the outlook in their markets and actions of their competitors. They don’t go out and, say, launch a new product because the Fed cut interest rates by half a point. The effect of central bank interest rate action is asymmetrical: it is much more effective in curtailing growth than stimulating it.
Having said that, there are some enterprises that do become more active when interest rates fall: leveraged speculators like private equity, hedge funds….and real estate developers and investors who borrow heavily. But the IMF ascertained that more financialization beyond a low level reduces growth. and asset management, like private equity and hedge funds, creates the biggest drag.
By Yao Chen, Associate Professor Erasmus School of Economics, Erasmus University Rotterdam, Felix Ward, Associate Professor, Erasmus School of Economics Erasmus University Rotterdam, and Florian van Genderen, PhD candidate Erasmus University Rotterdam. Originally published at VoxEU
Politicians have long pressed central banks for lower interest rates in the belief that easy money wins elections. This column tests that premise using 385 parliamentary and presidential elections in 18 countries between 1900 and 2020. It finds no evidence that easy money helps incumbents. The most likely electoral return to a typical easing is close to zero, and large returns are ruled out. A modest return can be decisive only in exceptionally close races. The electoral case for pressing central banks is weak.
Politicians have frequently sought easier monetary policy from their central banks. In recent years, President Trump has repeatedly demanded lower interest rates, and his attempt in 2025 to remove a Federal Reserve governor reached the Supreme Court, which blocked it in Trump v. Cook (2026).1There are notable historical precedents. Lyndon Johnson confronted the Federal Reserve chairman, William McChesney Martin, over a rate increase in 1965 (Meltzer 2009), and Richard Nixon pressed Martin’s successor, Arthur Burns, for easy money before the 1972 election (Abrams 2006). Those who lost office have sometimes blamed their central bank. George H. W. Bush, defeated in 1992, later said of the Federal Reserve chairman: “Alan Greenspan disappointed me. I think if the interest rates had been lowered more dramatically, that I would have been re-elected president.”2Such episodes reflect a widely held belief: that easier money helps governments win elections.
But does it? The answer matters because political pressure on central banks is common (Eijffinger and de Haan 2026) and costly: it lowers expected policy rates but raises expected long-term yields (Kind et al. 2020, Eichengreen et al. 2025), and it raises inflation and inflation expectations without raising output (Drechsel 2024). Yet if easy money wins elections, governments have a lasting incentive to press their central banks, even at a cost to the economy. If it does not, that incentive largely disappears. In a recent paper (Chen et al. 2026), we provide evidence on this question. Drawing on 385 parliamentary and presidential elections in 18 countries between 1900 and 2020, we estimate how exogenous changes in interest rates affect the vote share of incumbent governments.
Why Politicians May Be Wrong
Economic theory gives reasons to doubt the premise. Modern central banks set interest rates on the advice of large staffs of professional economists, so a government that overrides them must do better than that judgement in the eyes of voters. And forward-looking voters have little reason to reward stimulus that will raise inflation once the election has passed (Alesina 1987, Persson and Tabellini 2000). Only if voters are myopic, as early political business cycle models assumed (Nordhaus 1975), can easy money reliably buy votes.
The Challenge: Measuring the Effect of Easy Money on Votes
Measuring how interest rates affect the vote share of incumbent governments is hard. Central banks cut rates when the economy weakens, and a weak economy costs incumbents votes (Fair 1978, Lewis-Beck and Stegmaier 2000). A simple comparison of interest rates and election results therefore confounds the effect of monetary policy with that of the economic conditions to which it responds. Identifying the causal effect requires rate changes that are not driven by the domestic economy.
The trilemma of international finance supplies the right kind of variation (Jordà et al. 2020). A country that pegs its currency to a base country’s and allows capital to move freely must match the base country’s interest rate. Its rates then move with conditions abroad, not with its own economy or politics. An imported rate change thus does what successful political pressure on a central bank would do: it moves monetary policy off the course the central bank would have set for its own economy.3
Britain in 1992 is a case in point. The British pound had joined the Exchange Rate Mechanism of the European Monetary System in October 1990, tying it to the Deutsche Mark. After German reunification, the Bundesbank kept its rates high to contain inflation in Germany. Britain, with output stagnant after a recession and unemployment still rising, could not set rates for its own conditions. The Bank of England held its base rate at 10.5% through the spring of 1992, because a further cut would have put sterling’s parity at risk unless German rates fell first (Bank of England, 1992). It was under these imported rates that John Major’s Conservative government faced the voters on 9 April 1992. The government was returned to office with almost the same vote share as five years earlier, 41.9% against 42.2% in 1987.
Easy Money Does Not Buy Votes
This experience is typical. Across our sample, we find no evidence that easy money helps incumbents. For a typical easing of 0.91 percentage points over the term, our point estimate implies no gain at all. The data rule out large effects. At the upper bound of our confidence interval, a typical easing raises the incumbent’s vote share by 0.90 percentage points. Incumbents in our sample lose a median of 2.50 percentage points over a term, the cost of ruling (Nannestad and Paldam 2002). Even at the bound, offsetting that loss would take an easing larger than 99% of the exogenous easings observed in the sample.
Small Effects, but They Could Matter in Close Races
A null result is only convincing if it holds up under different assumptions. We therefore vary the choices on which reasonable analysts might differ. These include whether voters judge monetary policy over the whole term or only by the year before the election (Healy and Lenz 2014), whether the effect depends on the economy being in recession or inflation being high, and whether rate hikes and cuts have symmetric effects. In none of these variations does a typical easing move the incumbent’s vote share by more than 0.7 percentage points, and none of the estimates is statistically different from zero. The paper reports these checks in full.
None of these variations, however, rules out modest effects, and such effects can matter where elections are close. Our sample contains 104 elections in which an incumbent lost office, either a parliamentary majority or a presidency. At our point estimate, a typical easing would have reversed none of these losses. At the upper bound, it would have reversed 15, and across specifications between 5 and 19.
Figure 1 shows the closest calls. The closest is the US presidential election of 2000, decided in Florida by 537 votes. The next closest are parliamentary majorities lost by a few seats and presidencies lost narrowly in the decisive state or round. Most of the 104 losses fell short by far more than any bound could make up. What the data leave open is a modest effect that decides the occasional close election.
Figure 1 Close calls: Lost elections and the vote-share gain needed to keep office
Notes: Lost parliamentary majorities (dots) and presidencies (squares), ranked by the vote-share gain the incumbent needed to keep office. Shaded band: range of upper bounds across specifications on the gain from a typical easing (0.91 percentage points over the term). *Outside the estimation sample.
Source: Chen et al. (2026).
Conclusion
Taken together, our results amount to a bounded null. The most likely electoral return to easy money is close to zero, a large return is ruled out, and a modest return can be decisive only in an exceptionally close race. The evidence offers little support for the longstanding belief that easier money is an effective electoral tool.
This matters for the current debate. Pressure on the central bank has documented costs. Our results show how little easier money buys politicians in return. Over the past century, voters have judged governments on many things, but easier money has rarely been decisive at the ballot box.
__________
