Skip to main content

You are here

Blog Listing Page

The US yield curve is not a broken recession indicator

By Robin Marshall, director of fixed income research

The recent inversion of various parts of the US yield curve has revived concerns that this inversion is—as night follows day—a definitive recession omen. More recently, it has been suggested that the relationship between yield curve inversion and recession is obsolete.

The assertion that an inverted yield curve—presumably 10-year vs. 2-year yields—is a “broken barometer" of recession risks appears to be based on two propositions. First, the view that the Fed’s post-2009 quantitative easing programs have distorted the yield curve, rendering it flatter relative to previous cycles. But given how reliable the US yield curve has been in forecasting previous US recessions—inverting with a lead time of 12-18 months before each of the last eight US recessions in 10-year/2-year—ignoring a full yield curve inversion as a US recession indicator would require compelling counter evidence. 

Also, if a full yield curve inversion is defined to include 30-year/2-year inverting as well as 10-year/2-year and 5-year/2-year, all US recessions since the 1960s have followed within 12-18 months of a full yield curve inversion. More specifically, the widely quoted “false signal” of 1998 was only a partial yield curve inversion, since the 30-year/2-year did not invert. In this current cycle, the 30-year/2-year area has not yet inverted, though the difference in yield is now only 42 basis points.

There are technical reasons for thinking the 10-year/2-year section of the yield curve may be flatter in this cycle, mainly because the Fed is not actively managing the duration risk of its mortgage-backed securities portfolio. But the 5-year/2-year section of the yield curve has also inverted, which is a “cleaner" measure of recession risks, and normally inverts in advance of the 10-year/2-year, and 30-year/2-year (see Chart 1).

Chart 1: US yield curve and previous US recessions

Source: Datastream as of May 10, 2019. Past performance is no guarantee of future results. Please see the end for important legal disclosures.

The second suggestion is that we are seeing a replay of the late 1990s' Asia Financial Crisis, when there was a "false signal" from the US yield curve, but there appear to be few obvious similarities between the current cycle and when that so-called “false signal" occurred. Growth in the US was much stronger back in 1998-99—nominal GDP grew at 6.9% when the Fed first eased in Q3 1998 and the Greenspan Fed reduced interest rates following the collapse of Long Term Capital Management (LTCM) and the Asian financial shock—compared to 3.84% in Q1 2019. And bond yields were also much higher back then, with nominal and real yields at 5.5% and 3.5% in the 10-year area before the LTCM and Asian shocks compared to the current 2.42% nominal and 0.54% real yields.

The reasons behind the Fed policy shift are quite different. In 1998-99, the Fed eased by 75 basis points in response to perceived systemic risk that never materialized. In 2018-19, the Fed has pivoted to neutral policy in response to growth headwinds and persistent low inflation. The US economy has also been in a lower nominal GDP growth regime since 2009 (see Chart 2) compared to the late-1990s. And US real and interest rates have been falling steadily since October 2018, whereas they barely fell at all in 1998-99.

Chart 2 US nominal GDP growth and 10 year US Treasury yield

Source: Datastream as of May 10, 2019. Past performance is no guarantee of future results. Please see the end for important legal disclosures.

Canadian evidence casts doubt on view QE has been the main driver of US yield curve

Canadian yield curve evidence is also revealing. It is well known the Canadian economy is highly correlated with the US, because 75% of Canadian exports go to the US, so monetary policy settings are often similar. However, unlike the US, although the Bank of Canada (BoC) did supply "exceptional liquidity" to the financial system through term loan facilities, the BoC did not undertake QE purchases of government bonds during and after the GFC. So if QE purchases suppressed US Treasury yields and flattened the US yield curve, it should not have distorted or flattened the Canadian yield curve. At the very most, there might have been a gravitational pull from lower US Treasury yields.

But the Canadian yield curve has, indeed, inverted in 5-year/2-year, like the US in 2018-19. In addition, like other dollar-bloc alternatives to US Treasuries, Canadian government bonds now trade on lower yields than the US (see Chart 3). So Canadian evidence shows yield curves can invert without QE purchases, and outright yields decline for other reasons (low inflation, declining term premium, much lower nominal GDP growth regime, etc).

Chart 3: Canadian 10 year yields versus US Treasury 10 year yields

Source: FTSE Russell as of May 10 2019. Past data is no indication of future events. Please see the end for important legal disclosures.

This suggests QE may not have been the main driver of the flattening in the US yield curve, and that it may be unwise to dismiss the US yield curve as a recession forecaster, particularly if the 30-year/2-year section of the US curve follows the signal from the shorter maturities.



© 2019 London Stock Exchange Group plc and its applicable group undertakings (the “LSE Group”). The LSE Group includes (1) FTSE International Limited (“FTSE”), (2) Frank Russell Company (“Russell”), (3) FTSE Global Debt Capital Markets Inc. and FTSE Global Debt Capital Markets Limited (together, “FTSE Canada”), (4) MTSNext Limited (“MTSNext”), (5) Mergent, Inc. (“Mergent”), (6) FTSE Fixed Income LLC (“FTSE FI”) and (7) The Yield Book Inc (“YB”). All rights reserved.

FTSE Russell® is a trading name of FTSE, Russell, FTSE Canada, MTSNext, Mergent, FTSE FI, YB. “FTSE®”, “Russell®”, “FTSE Russell®”, “MTS®”, “FTSE4Good®”, “ICB®”, “Mergent®”, “The Yield Book®”  and all other trademarks and service marks used herein (whether registered or unregistered) are trademarks and/or service marks owned or licensed by the applicable member of the LSE Group or their respective licensors and are owned, or used under licence, by FTSE, Russell, MTSNext, FTSE Canada, Mergent,  FTSE FI, YB.  FTSE International Limited is authorised and regulated by the Financial Conduct Authority as a benchmark administrator.

All information is provided for information purposes only. All information and data contained in this publication is obtained by the LSE Group, from sources believed by it to be accurate and reliable. Because of the possibility of human and mechanical error as well as other factors, however, such information and data is provided "as is" without warranty of any kind. No member of the LSE Group nor their respective directors, officers, employees, partners or licensors make any claim, prediction, warranty or representation whatsoever, expressly or impliedly, either as to the accuracy, timeliness, completeness, merchantability of any information or of results to be obtained from the use of FTSE Russell indexes or research or the fitness or suitability of the FTSE Russell indexes or research for any particular purpose to which they might be put. Any representation of historical data accessible through FTSE Russell indexes or research is provided for information purposes only and is not a reliable indicator of future performance.

No responsibility or liability can be accepted by any member of the LSE Group nor their respective directors, officers, employees, partners or licensors for (a) any loss or damage in whole or in part caused by, resulting from, or relating to any error (negligent or otherwise) or other circumstance involved in procuring, collecting, compiling, interpreting, analysing, editing, transcribing, transmitting, communicating or delivering any such information or data or from use of this blog or links to this blog or (b) any direct, indirect, special, consequential or incidental damages whatsoever, even if any member of the LSE Group is advised in advance of the possibility of such damages, resulting from the use of, or inability to use, such information.

No member of the LSE Group nor their respective directors, officers, employees, partners or licensors provide investment advice and nothing contained in this blog or accessible through FTSE Russell indexes or research, including statistical data and industry reports, should be taken as constituting financial or investment advice or a financial promotion.

Past performance is no guarantee of future results. Charts and graphs are provided for illustrative purposes only. Index returns shown may not represent the results of the actual trading of investable assets. Certain returns shown may reflect back-tested performance. All performance presented prior to the index inception date is back-tested performance. Back-tested performance is not actual performance, but is hypothetical. The back-test calculations are based on the same methodology that was in effect when the index was officially launched. However, back- tested data may reflect the application of the index methodology with the benefit of hindsight, and the historic calculations of an index may change from month to month based on revisions to the underlying economic data used in the calculation of the index.

This publication may contain forward-looking assessments. These are based upon a number of assumptions concerning future conditions that ultimately may prove to be inaccurate. Such forward-looking assessments are subject to risks and uncertainties and may be affected by various factors that may cause actual results to differ materially. No member of the LSE Group nor their licensors assume any duty to and do not undertake to update forward-looking assessments.

No part of this information may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the applicable member of the LSE Group. Use and distribution of the LSE Group data requires a licence from FTSE, Russell, FTSE Canada, MTSNext, Mergent, FTSE FI, YB, and/or their respective licensors.

Blog Listing Page