Policy paper

Annex: The impact of borrowing on interest rates

Published 1 December 2023

An unanticipated increase in spending, or reduction in taxation, that is funded by additional government borrowing, will increase the level of demand in the economy, thereby increasing inflationary pressures, which may lead to an inflation-targeting central bank increasing interest rates ^{[footnote 1]}.

Economists have proposed various methods to estimate how much interest rates would increase in response to a given level of additional borrowing (see the table below for some examples and their results). HM Treasury has conducted some analysis on this question based on an application of the published OBR simplified macroeconomic model (the ‘small model’ ^{[footnote 2]}), a calibrated New Keynesian model of the UK economy.

A version of the small model has been previously used by the OBR alongside its main macroeconomic model to model the impact of unanticipated changes in fiscal policy on a small set of macroeconomic variables of interest, including interest rates. For fiscal events, the OBR usually conditions its forecast on the assumption that fiscal policy changes are anticipated and are therefore already ‘priced’ into market expectations of Bank Rate.

The small model is based on 4 key economic variables: the output gap, the inflation rate, Bank Rate and the (trade-weighted) nominal exchange rate. The relationship between these variables is determined by a system of key macroeconomic equations: the investment-saving (IS) curve; the ‘Taylor rule’; the Phillips curve; and an uncovered interest parity relation ^{[footnote 3]}. It should be noted that this approach does not consider potential supply-side benefits of certain fiscal policies. That means the impact on interest rates would be smaller where those policies do have a material benefit for the supply side. The approach also models the behaviour of the MPC in a stylised, rule-based way: in practice, the MPC will make interest rate decisions based on the full range of specific factors it would be faced with ^{[footnote 4]}.

To model the effect of an increase in borrowing, first its elements are specified so it creates a ‘demand shock’ that can then be used as an input to the small model. This is done by assessing the impact of a 1-year, 1% of GDP (£25 billion ^{[footnote 5]}) fiscal policy loosening, assuming proportionate contributions from RDEL, CDEL, AME and tax to PSNB. Subsequently, the OBR’s standard fiscal multipliers ^{[footnote 6]} are used to calculate the impact on aggregate demand. This results in a roughly 0.5% output gap shock.  The inflation and interest rate responses are then generated by the system of equations in the model.

This stylised shock generates a peak increase in interest rates of 60 basis points, resulting from a 1-year, 1% of GDP fiscal policy loosening as described above.

Greater-than-expected inflation persistence has led to suggestions from some economists that inflation and interest rates might be more sensitive now than usual to demand shocks. To assess this, scenario analysis is carried out, informed by recent contributions to the debate on why inflation is more persistent than expected. This entails:

• Increasing the sensitivity of inflation to changes in slack in the economy by increasing the slope of the Phillips curve (from 0.15 to 0.3).  This scenario captures the recent evidence that inflation may now be more sensitive to the level of slack in the economy, reflecting increased labour market tightness. This can be seen as the UK economy currently being on a steeper part of the Phillips curve than usual. Notable recent contributions to the literature on the Phillips curve aiming to explain the increase in inflation over the past two years have suggested that the Phillips curve has steepened, due to non-linearities arising from labour shortages and labour market tightness (based on evidence in the US; Benigno & Eggertsson, 2023), or possibly due to digitalisation and de-globalisation (IMF analysis of the UK and other advanced economies; Ari et al, 2023) ^{[footnote 7]}. In the last two years, estimation of UK inflation against measures of economic slack also suggests that there might have been a significant change in the relationship between inflation and unemployment with the implied Phillips curve slope increasing by at least a factor of 2, consistent with the UK labour market tightening significantly ^{[footnote 8]}. It should be noted that there is some emerging evidence of a loosening in labour market conditions in the very latest months.

• Increasing inflation persistence by increasing the coefficient on lagged inflation in the Phillips curve (from 0.7 to 0.9). This scenario is based on recent evidence linking greater-than-expected inflation persistence to agents’ price-setting behaviour being more backward-looking than usual. In speeches over the course of the last year, Catherine Mann has referenced analysis suggesting that the share of backward-looking agents has varied significantly over time and increases when energy prices surge ^{[footnote 9]}. In particular, she references work suggesting that the degree of backward-looking price-setting behaviour increases the longer that inflation spends away from target. In addition, the original OBR working paper references changes in inflation expectations as a potential source of inflation persistence and initially used a coefficient on lagged inflation of 0.9, which was later revised down to 0.7.

There is therefore a reasonable case that these updated coefficients provide a plausible description of how the UK economy may have been operating in recent times, having experienced, for example, a tight labour market and a steep rise in energy prices. These coefficients, and associated estimates, would be expected to return to more normal levels as inflation returns to target and slack emerges in the economy.

With these coefficients, a fiscal policy loosening of 1% of GDP could lead to a peak increase in interest rates of around 125 basis points.

Bringing these estimates together, a cautious range can be constructed by spanning across what the standard model coefficients would give us, through to what the scenario-based updated coefficients estimate. That suggests that an increase in borrowing of 1% of GDP (£25 billion ^{[footnote 10]}) might increase interest rates by between 50 and 125bps (to the nearest 25 bps). That range is higher than those found in the external literature (below); although they are broadly comparable, and relevant research on this topic almost always finds a significant and meaningful effect.

Published external estimates – macroeconomic impact from a 1% of GDP fiscal loosening

Detailed below are a range of interest rate impacts resulting from a fiscal loosening, whereby estimates relating to a fiscal tightening are assumed to also hold for a fiscal loosening. The models and estimation periods vary substantially, as do the duration of policy shock and timings of peak of impact across the studies considered.

IMF paper (July 2023) ^{[footnote 11]}

Area: Euro area
Impact on growth: NA
Impact on inflation: +0.15-0.25 ppts
Impact on interest rates: Higher between 30-50 bps

Caveats:

• Modelled impact of a fiscal tightening of 1% of GDP for 2 years and 0.5% of GDP in the third year, so impact likely to be weaker in comparison.
• HMT analysis assumes linearity in the model, with the results being symmetric (for example, holding for a fiscal loosening).
• Refers to impact on core inflation, impact on overall inflation could be stronger.
• The euro area evidence may not hold for the UK.

OBR small model paper (2012) ^{[footnote 12]}

Area: UK
Impact on growth: 0.5% higher output shock
Impact on inflation: +0.25ppts
Impact on interest rates: Higher by 50 bps

Caveats:

• HMT analysis assumes linearity in the model, with the results being scalable.
• It is not clear what the source of the demand shock is.
• This is a modelling exercise (calibrated on UK data), rather than an empirical study.
• The results are scaled to reflect the impact of a 0.5% output gap shock (closer to the HMT analysis), though the paper models a 1% output gap shock. The shock is identical in length to the HMT analysis (4 quarters).

IMF Blog on Fiscal Monitor (April 2023) ^{[footnote 13]}

Area: Advanced economies
Impact on growth: NA
Impact on inflation: +0.50 ppts
Impact on interest rates: Varies markedly

Caveats:

• HMT analysis assumes linearity in these estimates, with the results being symmetric (for example, holding for a fiscal loosening).

• Empirical estimates since 1985.

BIS paper (July 2022) ^{[footnote 14]}

Area: Advanced economies
Impact on growth: NA
Impact on inflation: +0.20 ppts
Impact on interest rates: NA

Caveats:

• Empirically estimated impact of a one percentage point increase in the fiscal deficit on the annualised inflation rate over the next two years, from 1972 onwards.

1. See Cloyne (2013) for an empirical estimation of the impact of fiscal policy shocks in the UK, and for its impact on interest rates, and Ramey (2019) for a broader literature review of the impact of fiscal policy on the economy. ↩

2. Working Paper No.4, ‘A small model of the UK economy’, Office for Budget Responsibility, July 2012. This HMT analysis is based on an updated version of the model described in the paper. ↩

3. A further equation captures changes in the cyclical component of the government’s primary balance. ↩

4. In 2017, Mark Carney, then Governor of the Bank of England, explained that “taking the past as a guide to the future is to assume that the nature [of] shocks does not change and that the economy remains resilient” when explaining that models or monetary policy rule-based frameworks should not be normative prescriptions (guides for what the central bank should do in all circumstances) for central bank decisions. (Mark Carney (2017) ‘Lambda’ speech). ↩

5. HM Treasury calculations to the nearest £1 billion in 2022 prices (the first full year for which data is available) using GDP first quarterly estimate, UK: July to September 2023 (series ID: BKTL), Office for National Statistics ↩

6. Economic and fiscal outlook, Office for Budget Responsibility, March 2023, Box 2.2: The economic effects of policy measures. ↩

7. Benigno & Eggertsson (2023), “It’s Baaack: The Surge in Inflation in the 2020s and the Return of the Non-Linear Phillips Curve”; Ari et al (2023), “Has the Phillips Curve Become Steeper?” ↩

8. Members of the Monetary Policy Committee have referenced this possible steepening in the Phillips curve in the last 2 years. See, for example,  Huw Pill’s speech, ‘Recent developments in UK monetary policy’, in August 2023 and Catherine Mann’s September 2022 speech, ‘Inflation expectations, inflation persistence, and monetary policy strategy’. ↩

9. Cornea-Madeira and Madeira (2022), referenced in Catherine Mann’s February 2023 speech on the role of inflation expectations on the transmission of monetary policy, find that the share of backward-looking agents varies significantly over time, and is higher when energy prices increase. See the appendix for more detail. Catherine Mann also references backward-looking price setting behaviour in her September 2022 speech. ↩

10. HM Treasury calculations to the nearest £1 billion in 2022 prices (the first full year for which data is available) using GDP first quarterly estimate, UK: July to September 2023 (series ID: BKTL), Office for National Statistics ↩

11. Shared Problem, Shared Solution: Benefits from Fiscal-Monetary Interactions in the Euro Area (imf.org) ↩

12. Working-paper-No4-A-small-model-of-the-UK-economy.pdf (obr.uk) ↩

13. Fiscal Policy Can Help Tame Inflation and Protect the Most Vulnerable (imf.org); or Chapter 2 of the Fiscal Monitor Fiscal Monitor April 2023 (imf.org) ↩

14. Fiscal deficits and inflation risks: the role of fiscal and monetary regimes (bis.org) ↩