The US inflation outlook and implications for Fed policy
With markets expecting even less cash rate rises than the Fed, even a modest acceleration of inflation could result in an upward adjustment for bond yields and increased volatility in equity markets.
The US Federal Reserve (Fed) raised US cash rates in December, the first rate hike in almost 10 years, with expectations of a gradual continuation, it has pencilled in a further 1.0% over 2016. While their unemployment target has been largely met, modest expectations of policy have been able to be maintained as US inflation has remained benign. As long as inflation remains in line with the Fed’s expectation of a rise to its 2% p.a. target, the Fed will be able to continue along its expected gradual path. However, with markets expecting even less cash rate rises than the Fed, even a modest acceleration of inflation could result in an upward adjustment for bond yields and increased volatility in equity markets.
We believe the risk to this scenario is a pickup in inflation to above the Fed’s 2% target, due to a strengthening global economic environment, in response to the recent surge in global money growth. This would see the perception that the Fed is “behind the curve” and lead to an acceleration in the tightening process. With the Fed moving from taking the foot off the accelerator in a slow and controlled manner to tapping on the breaks, it could see considerable damage to bond markets, with bond yields expected to adjust significantly higher, reflecting the higher inflation environment. For equity markets the initial shock would see an increase in volatility and a re-pricing of equity markets to a higher yield environment. But then equity markets may perform reasonably due to a pick-up in profits, until signs that Fed action is impacting on the growth outlook.
The US inflation outlook is critical in our view to the size and pace of future Fed tightening. In this paper we explore the short term inflation dynamics, while previously we have looked at the medium term inflation outlook (see Stevenson, 2013). Two models are built to forecast the quarterly changes in the core personal consumption expenditure (PCE) deflator. The PCE is used over the CPI as this is the Fed’s preferred measure of inflation. They argue it takes into account changing consumer preferences better than the fixed weight CPI.
The two models developed were: a simple statistical model , and a Phillips Curve model . Both models show the importance of oil prices in the path of core inflation in the US, driven by the second round impacts of changes in oil prices and as a proxy for global demand. The analysis finds that a stabilisation in oil prices and the US dollar should see the core PCE deflator head towards the Fed’s 2% target. This supports the Fed’s expected path in the cash rate - moving toward neutral in a controlled manner. However, a re-synchronisation or acceleration of global growth would pose a risk to this outlook and would likely lead to a more aggressive policy path. With the bond market pricing in neither scenario, we believe their remains significant upside to bond yields.
Using a simple statistical model
Initially, forecasting exercises were conducted based on an enhancement of a univariate model developed by Zaman (2013). He presented a simple technique that significantly improved the forecast accuracy of a univariate inflation model, two to three years out. Rather than forecast inflation by modelling it on past values of itself, the model is based on, what he called the PCE inflation gap, or the annualised quarterly inflation rate minus long run inflation expectations. This allows the model to take into account the varying medium term trend in inflation, i.e. the high inflation in the 70s and 80s as opposed to recent outcomes. This creates what is called a stationary series (a constant mean), which leads to a 20 to 30 percent gain in forecast accuracy.
Figure 1: PCE Inflation and PCE Inflation Gap
Source: Schroders, Zaman (2013). Shaded areas indictate recessions.
One thing you can notice when considering the PCE inflation gap is that large deviations from the mean are consistent with periods of large movements in oil prices. This is not surprising given Zaman modelled headline inflation, but of interest is that this is also seen when calculating the PCE inflation gap based on the core PCE deflator.
To take advantage of this relationship a two variable Vector Autoregressive (VAR) model was built. The two variables; core PCE gap and oil prices; were regressed on past lags of themselves and past lags of the other variable. That is, the core PCE gap was regressed on the previous four quarters of itself and the past four quarterly changes in oil prices, while the quarterly change in the oil prices was regressed on the previous four quarters of itself and the past four quarters of the core PCE gap. The statistical strength of the core PCE model is strong, with an adjusted R squared of 0.59. While, not surprisingly, the strength of the oil price model was weaker – adjusted R squared of 0.15.
Figure 2: Core PCE Inflation and Forecast – Statistical Model
The VAR model can be used to forecast the core PCE deflator and the results can be seen in figure 2. The model suggests core inflation will lift slightly and grow over 2017 at a rate of 1.4% p.a., seeing core inflation remaining below the Fed’s target for the foreseeable future, an environment where the Fed can maintain its gradual approach. Another approach of taking the oil price assumptions from the futures market (a gradual rise to $52pb in December 2017) leads to a similar answer.
Using a fundamental model
The advantages of the model above are that it is simple and relatively robust. However, while capturing supply side shocks, it sees oil prices proxy for several factors, especially US and global growth. Generally this is not a problem given the high correlation between the US and global economies, but with the US economy stronger than the rest of the world, this may understate the inflation risk. To develop this further a model based on work by Goldman Sachs (Hatzius, 2015) was built. This is a Philips Curve based model, seeing the core PCE deflator a function of labour market utilisation (measured by U6 unemployment rate), energy prices (year on year change in consumer energy prices, current and lagged 4 quarters), and currency (quarter on quarter change in USD TWI, current and lagged one quarter).
The model was estimated from 1997, given this is the period where inflation expectations have been well anchored (see figure 1). Inflation regimes have had an impact on the Phillips Curve and it has become flatter over time, labour market utilisation is having a lower impact on inflation dynamics, as inflation expectations have fallen and become anchored. Models built on longer time frames will overstate the impact of labour market utilisation on inflationary dynamics. This may be due to globalisation, seeing the oil price variable capture global demand and offsetting some of the power of the domestic labour market.
Figure 3: Core PCE and Forecast – Phillips Curve Model
Figure 3 shows the fit of the model (R squared 0.40) and forecasts based on the last 5 years downward trend in the U6 unemployment rate continuing, the expected path in oil prices from the futures market (a gradual rise to $52pb in December 2017), and a relatively flat USD TWI based on the futures market. Not surprisingly, the Phillips Curve model forecasts a higher inflation outcome than the statistical model as it captures the stronger nature of the US economy. Another take away from the model is that core inflation is currently low due to the very large falls in the oil price and the rise in the US dollar.
Testing the sensitivity of the model we ran through a couple of scenarios. First, if the WTI oil price falls to $15 over 2016 (1998’s $10 low adjusted for inflation), and the currency maintains its current inverse relationship, the model forecasts core PCE inflation of 1.0% p.a. over 2016. On the up side, if oil prices rise back to $100 over 2016, with falls in the US dollar, the model forecasts core PCE inflation at the Fed’s target, over 2016, and higher in 2017 as oil price rises flow through with a lag.
Outlook for inflation
The above modelling suggests for core PCE inflation to remain this low, a continued rapid fall in oil prices is required, as it is the change in oil prices that feeds through into the formation of inflation. A stabilisation would see the suppression of inflation from the falling oil price dissipate and see it head back towards the Fed’s 2% target (approximately 2.5% for the core CPI).
Both models are currently undershooting the actual inflation rate, both suggesting a 1% inflation rate as opposed to reported 1.3%, when inflation usually overshoots the model forecasts on the up and down-side. This current differential between actual and forecast is most likely related to bottom up pressures from housing and healthcare sectors, which are expected to continue in the near term. Taking this into account our central case is for core inflation to rise to the Fed’s target of 2% over the next two years. This is consistent with the cyclical component of the Australian Fixed Income and Multi-Asset team’s investment process. Where the US economic cycle is seen to be in the mid-cycle stage – a positive output gap (rising inflationary pressures) and a positive growth gap (growth above trend).
The recent weakness in oil prices has been driven by both a belated pricing in of the new supply environment (with the rise of fracking) and the continued weakness in the global economic environment (impacting on the demand for oil). With the market most likely to have adjusted to the new supply dynamic, barring short term momentum, oil prices will likely be driven by the outlook of the global economy. The recent surge in global narrow money, led by China and the Eurozone, suggests the risk is to the upside of global growth and rising oil price. This would see core inflation rising to above the Fed’s 2% target and a much more aggressive policy response. While this is not our base case, we believe it is of significant probability to watch closely for signs of acceleration in global growth.
Figure 4: Global Money Growth and GDP
The base case scenario of a rise in core PCE inflation to 2.0% p.a. over the next couple of years is consistent with the current Fed views and therefore consistent with their current policy expectations – approximately 1% lift in the official cash rate each year over the next three years. Policy expectations implied from markets are a fraction of this, with the cash rate expected to be around 1.5% at the end of 2018, opposed to the Fed’s expectations of 3.25%. This suggests vulnerability to bond markets as markets adjust to a more aggressive than expected policy environment. This would be a difficult environment for equity markets, rising rates generally lead to a downward rating via lower PE ratios. Usually this is more than offset by rising profits. However, with profit margins elevated and moderating somewhat, profitability has been poor - this may not be the case.
Of course if global growth responds to the surge in narrow money, then it is a different situation. The Fed will be behind the curve with inflation expected to run above its target. This would see a large gap between current market pricing and inflation fundamentals, and a significant upward adjustment would be expected in bond yields. For equity markets it gets trickier: lower PE in a rising interest rate environment, but much better profits due to improved economic growth. Most likely, the initial shock would see an increase in volatility, but then equity markets would perform well until signs that Fed action is impacting on the growth outlook. However, our overall view on the medium term outlook for equities is modest, based on valuation metrics. With below average returns, in the low to mid- single digits, from global developed markets, and slightly below average returns in Australia, this is due to our more neutral valuation position.
Hatzius, J. (2015), “The Risks to Inflation, and to Liftoff in 2015”, Goldman Sachs US Daily, 7 January.
Stevenson, S. (2013), “Forecasting Inflation”, Schroders Talking Point, November.
Zaman, S. (2013), “Improving Inflation Forecasts in the Medium to Long Term”, Federal Reserve Bank of Cleveland, Economic Commentary, no. 2013-16.
1 There are several differences between the two series but the main ones are: formula effects and weight effects. Formula differences are related to the CPI being fixed weight while the PCE takes into account consumer substitution as relative price changes (e.g. rising beef prices may lead to less beef consumed and more chicken). Weighting differences are driven by the data source, the CPI from household surveys while the PCE is from business surveys. The CPI covers households out of pocket expenses, while the PCE also covers expenses by employers or federal programs on behalf of consumers (e.g. medical insurance payments are in the PCE but not the CPI). The core CPI has on average run 0.5% p.a. higher than the core PCE deflator.
2 The model is not based on economic theory but statistical correlation between inflation, past values of itself, and past values of changes in oil prices.
3 Phillips Curve models are a theoretical approach that links inflation to labour market slack and supply side shocks.
4 A univariate model - in this case inflation is modelled on past values of itself. These models generally perform well, in and out of sample, and are a good benchmark for fundamental models.
Opinions, estimates and projections in this article constitute the current judgement of the author as of the date of this article. They do not necessarily reflect the opinions of Schroder Investment Management Australia Limited, ABN 22 000 443 274, AFS Licence 226473 ("Schroders") or any member of the Schroders Group and are subject to change without notice. In preparing this document, we have relied upon and assumed, without independent verification, the accuracy and completeness of all information available from public sources or which was otherwise reviewed by us. Schroders does not give any warranty as to the accuracy, reliability or completeness of information which is contained in this article. Except insofar as liability under any statute cannot be excluded, Schroders and its directors, employees, consultants or any company in the Schroders Group do not accept any liability (whether arising in contract, in tort or negligence or otherwise) for any error or omission in this article or for any resulting loss or damage (whether direct, indirect, consequential or otherwise) suffered by the recipient of this article or any other person. This document does not contain, and should not be relied on as containing any investment, accounting, legal or tax advice. Schroders may record and monitor telephone calls for security, training and compliance purposes.