![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
ABSTRACT
The ‘power of bond markets’ is a widely assumed and poorly understood feature of the global economy. We demonstrate that even in a bond market as stable as the United States this influence is considerable. In this article, we scrutinise a particularly direct influence, the impact of US Treasury yields on presidential approval rates. Our empirical analysis from 1961 to 2010 demonstrates that rising/falling bond yields lead to a decline/increase in approval rates. We show that this impact is mediated via the US mortgage market. The stronger the rise in mortgage rates, the stronger the influence of Treasury yields on presidential approval. We then outline the broader possible political impacts of this, particularly given foreign and domestic central bank ownership of US Treasuries.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes on contributors
Iain Hardie is Senior Lecturer in International Relations and the University of Edinburgh. He has published in journals including World Politics, New Political Economy, Review of International Political Economy and Socio-Economic Review. His most recent co-authored book, Chains of Finance, is published by Oxford University Press.
Ailsa Henderson is Professor of Political Science at the University of Edinburgh where she conducts research on comparative political behaviour. Recent articles have appeared in Journal of European Public Policy, Post-Soviet Affairs, Publius, Electoral Studies, and British Journal of Politics and International Relations. Her most recent co-authored book, The National Question and Electoral Politics in Quebec and Scotland, is published by McGill-Queen’s University Press.
Charlotte Rommerskirchen is Senior Lecturer in International Political Economy at the University of Edinburgh. Her research interests lie in sovereign debt management, the politics of crises and financial globalisation. Her first monograph, Free Riders on the Storm: EU Fiscal Policy Coordination in Hard Times, has been published by Oxford University Press.
Notes
1 The market yields which determine the cost of government borrowing of course represent the aggregation of the actions of market actors in reacting to information; the degree to which yields rise or fall as a result of any president’s policies or changing macroeconomic conditions will be determined by buying and selling of US Treasuries that result. Although external events will have an impact on yields, therefore, this is only because of investor assessment of those events, and this assessment will vary by borrower (Mosley Citation2003), investor (Hardie Citation2012) or indeed over time.
2 These patrimonial explanations are what Piketty (Citation2018) has recently referred to as the Brahmin Left and Merchant Right, or, in one summary the ‘haves and have yachts’ (Kuper Citation2018)
3 For recent research on the impact of market pressures on politics see inter alia Mosley Citation2003: esp. ch.5; Hardie Citation2012; Campello Citation2015.
4 Lewis-Beck and Stegmaier (Citation2000), in their review of the literature, conclude that evidence for retrospective pocket book voting is low, but that voters’ prospective views on their personal economic situation have a significant influence.
5 Sovereign bond yields measure the interest payments that investors demand to lend money to governments. This domestic-sided description does not take into account external factors, notably international risk factors (so-called push factors). Our quantitative models account for global risk aversion.
6 In one estimate, annual volumes varied from US$2532 million to US$234 billion for the period 2000–2016, compared to variation in borrowing for home purchase of US$1512 and US$505 billion (Mortgage Bankers Association Citation2015, p. 10).
7 There are also ways in which higher government bond yields help individuals financially, but these are less prevalent. For example, pensioners can purchase a fixed-rate annuity to give them a guaranteed (nominal) income. The income they receive will be higher if government bond yields are higher, as insurers use bond yields to price new annuities (similar to the pricing of mortgages, US bond yields are used as a benchmark rate for the domestic annuity market). The US annuity market, relative to the size of its economy and mortgage market, is small. ‘Given the choice, people do not choose to annuitise as expected to when attaining the end of their working lives’ (Rusconi Citation2008). Along similar lines, pension plans rely on high bond yields to hit their target investment returns. The projections of future retirement income will likely to lower and the deficit on salary-linked pension schemes higher when bond yields are low. Persistently low bond yields therefore may put pensioner’s payouts at risk. Although this poses a real threat to the sustainability of pension plans, its materialisation is not likely to be captured in the time-frame of our analysis when most pensions have been considered relatively safe and the underfunding of existing schemes cannot be solely attributed to a low-yield environment.
8 The widely accepted ‘money illusion’ suggests that individuals focus more on nominal than real monetary values (see, e.g. Shafir et al. Citation1997).
9 Serious economic difficulties may well also result in falling inflation.
10 The so-called honeymoon effect is not a new phenomenon: Thomas Jefferson (Citation1796) famously proclaimed that ‘I know well that no man will ever bring out of that office the reputation which carries him into it. The honey moon would be as short in that case as in any other, and its moments of ecstasy would be ransomed by years of torment and hatred’.
11 A dummy variable controlling for the (failed) impeachment procedure against President Clinton in the aftermath of the Lewinsky affair fails to reach statistical significance (and is not retained in the final model). This confirms Zaller’s (Citation1998) and Newman’s (Citation2002) finding that although approval models show that the public punishes presidents for scandals, Clinton remained popular after several scandals, – indeed approval ratings actually increased during the Lewinsky investigation and impeachment proceedings.
12 Following Norpoth (Citation1984), we code the variable Vietnam as (−1) under Johnson and (+1) under Nixon (and (0) elsewhere). The effects of both Gulf Wars are measured by two dummies equal to (1) between August 1990 and January 1991 and between March and May of 2003. We also control for the patriotic revival after 9/11. The increase in presidential approval (from 55 per cent in August 2001 to 89 per cent in September 2001) is the most substantial boost yet recorded, overtaking FDR’s approval surge after Pearl Harbor. What is more this effect has been slower to decay than previous rallies (Gaines Citation2002, Hetherington and Nelson Citation2003). Therefore, instead of including a binary dummy variable, we create a variable that is zero in the quarters prior to 11 September, and 1/i starting from that quarter (with i = 1, 2, 3, …). In addition, we control for the effects of two scandals involving the president. First, Watergate is a dummy taking the value (1) from July 1973 to August 1974, and (0) otherwise. Second, Iran-Contra is a dummy equal to (1) between November 1986 and March 1987, and (0) otherwise.
13 There is no agreement on whether economic determinants enter the popularity function contemporaneously or with a time-lag (t- …). We also ran our model with lagged economic controls (t−1). Results hold (see Online Appendix Figure A2).
14 There is a wider debate as to the pros and cons of lagged dependent variables, which has been reviewed and analysed by Achen (Citation2000, see also Beck Citation1991, p. 66).
15 Using alternate lag values (either larger or smaller) had no effect on the overall results.
16 We also correlated the difference in 10-year bond yields with other housing market variables, namely house prices, homeownership rates and household debt and found no strong pattern of association.
17 We also ran our interaction model with lags of the change in bond yields (t−1, t−2, and t−3, respectively), where the marginal effect was statistically not significant. This suggests that the impact of a change in bond yields on presidential approval is immediate.
18 There seems to be however no simple correlation between presidential approval rates and the variables Mortgage, Mortage2 and ΔYield*Mortgage.
19 Source: US Treasury.
20 In parallel, some interpretations of modern monetary theory (MMT) argue that governments borrowing in their own currency face no hard budget constraint, as they can always create money to repay their debts, and therefore budget deficits do not matter. MMT scholars have denied making such a claim (e.g., Black Citation2019). The increasingly vituperative debate around MMT has involved accusations that some leading economists have misunderstood the theory, and we would not claim the expertise to engage in it. However, any policy move influenced by MMT involving the greater integration of monetary and fiscal policy and control of government bond yields would, it is suggested here, have direct implications for presidential approval. An alternative outcome of such debates, simply a more relaxed view of fiscal deficits, would only make the political implications identified here more important.
21 As of 15 August 2018. Source: FRED Economic Data, Federal Reserve Bank of St. Louis.