US - Anything But Normal, Part II

Can the US could break out of the ‘new normal’ patterns of sluggish acyclical growth seen since 2010? Part I showed that improving trends in return on capital and real labour productivity generate a background in which we should expect rising capital investment and rising employment. At the least, this suggests an uptick in GDP growth in the short and medium term. In addition, it looked at two potential subcyclicals that could accelerate that growth into a genuine business cycle upswing - ie, a breakout from the ‘new normal’.  The dynamics for both the capital goods sector and for inventory management - both of which can act as business cycle accelerators - have been in a ‘negative adjustment’ phase over the last two years , but this has left them, ironically, increasingly sensitive to the current rise in global producer prices. 

But if all these provide fuel for a potential business cycle upswing, could these nevertheless be snuffed out by a deterioration in saving/consumption decisions in the household sector and, more generally, a tightening of credit and monetary conditions?  

This piece makes three points: 

  1. The current tightening in monetary conditions are not dramatic, and pose no material threat to GDP growth in 2017/18.
  2. Bond yields have already risen further than is easily explained either by current supply/demand conditions in bond markets, or by what consensus expects of Fed rate rises, and the prospects for inflation and GDP over the coming five quarters. In the absence of a monetary policy ‘mistake’ bond markets alone are not about to snuff out a significant acceleration in economic growth.
  3. Nevertheless, household savings rates fell gently but persistently throughout 2016 and are now at historically low levels. A reversion to the mean would cut approximately 30bps off GDP growth. In a ‘new normal’ scenario, in which GDP averages 2.1% with a standard deviation of 70bps, this is a material risk. 

On balance, then, monetary conditions and bond markets are not currently threatening, and are unlikely to become so unless the Fed is blindsided by a stronger-than-expected inflationary outbreak. But the fall in the household savings rate does represent a material risk, underlining the importance of sustained consumer confidence and, in particular, confidence in labour markets. 

A noticeable GDP acceleration seems very likely, and a breakout sustained above 3% possible, but a sustained rise in savings rate is the likeliest reason why growth would be sustained at the ‘new normal’ upper bound of around 2.7%.

Monetary Conditions Have Tightened, But Not Much
At its broadest, monetary conditions tightened during 2H16. My monetary conditions indicator includes the four crucial factors about money: 
i) how much there is of it (ie, monetary aggregates); 
ii) how much it costs (ie, 10yr bond yields); 
iii) how much you are paid to save it rather than spend it (ie, the shape of the yield curve); and
iv) what’s happening to its international value (ie, movements vs the SDR). 

The two main factors generating the tightening have been:
i) the rise in the dollar, which rose 4.6% against the SDR between May 2016 and December 2016 (before stabilizing in January), and
ii) the slowdown in monetary growth, which was sufficiently gentle not to be reflected yet in yoy growth, but which set in in July and intensified all the way to December. During that time M2 growth actually accelerated from 6.7% yoy in a peak of 7.7% in October. By December the yoy rate had retreated to 7.1%, but it is unlikely for it to see 7% again any time soon in 2017. 

That slowdown in monetary growth is echoed by a sharper slowdown in bank lending, with total loan growth losing momentum very sharply in November and December, with monthly movements 1.6SDs and 1.8SDs below historic seasonal trend respectively. The slowdown was particularly marked for commercial and industrial loans, with bank’s loans outstanding essentially stagnant since mid-year: 

The negative impact of the strength of the dollar and the cramping of monetary growth and bank lending was partly offset by the fall in real bond yields. Nominal bond yields rose sharply but so did inflation, so by December 10yr bond yields of 2.5% were only approximately 40bps higher than the 2.1% yoy rise in CPI inflation. Since the beginning of 2011, real bond yields have averaged 61bps, so these were low real yields even by ‘new normal’ standards. . 

Although monetary conditions have tightened, recent history suggests the tightening is unlikely to be decisively depressing to the domestic economy. During the ‘new normal’ period, large and sustained fluctuations in monetary conditions have probably had some general and weak correlation with movements in final sales in real GDP (ie, GDP minus change in inventories). However, as the chart below suggests:
i) the relationship is weak. One has to take 12ma for both in order even to make a visual connection between the two, and even on this generous basis, the relationship would not generate anything statistically interesting;
ii) it confirms that the deterioration in monetary conditions seen in 2016 is currently so mild as to be undetectable on a 12m basis; and
iii) if one is to believe that some sort of relationship between the two exists, one would be looking for an acceleration in GDP in 2017 to answer the overall improvement in conditions seen throughout 2016! 


Bond Yields Have Got Ahead of the Game
Bond yields have risen more sharply than is justified by net bond issuance (ie, the gross issuance of non-financial bonds minus the change in bonds held by US banks). 

During the last six years (ie, during the ‘new normal’) changes in bond yields have coincided reasonably regularly with changes in this net bond issuance metric. But the jump in bond yields during 4Q was not accompanied by any significant rise in net bond issuance. Rather, net bond issuance has merely flattened out. If this relationship is maintained (and although the relationship makes good economic sense, there’s no guarantee it will be maintained) then it looks as if current bond yields are anticipating a significant rise in net issuance. And if that is the case, even if that net issuance materializes, the scope for further significant bond yield shocks is limited. 

This same message also emerges when one deploys the usual bond-yield models, which seek to explain movements in yields from movements in Fed Funds rate, inflation and GDP growth. The relevant current consensus forecasts for the coming five quarters include: four Fed rate increases, taking rates from the current 75bps to 1.45%; CPI hovering around 2.4%, and GDP steady at around 2.3%. Plugging these assumptions into these models suggests:

  • rise of 20-25bps between the low of 3Q16 and 1Q17, and a rise of 60-70bps between 3Q16 and 1Q18.  
  • In fact, what we’ve already seen is a rise of over 90bps since 3Q16, and consensus estimates expect that rise to stretch to 134bps by 1Q18. 

One should have only moderate faith in models and consensus estimates. Nonetheless, they do tend to suggest that bond markets have already discounted significant upward surprises in some or all of i) Fed policy tightening; ii) accelerating inflation and iii) sustained GDP growth. 

Perhaps bond markets do quietly expect a breakout from the ‘new normal’ after all, even if consensus economic forecasts do not. 

Household Savings Rates Have Fallen, Are Vulnerable to a Reversal
Neither the current state of monetary conditions, nor potential rises in bond yields look to pose a significant issue currently. But the drop in household savings seen during 2016 does generate a vulnerability to domestic demand in the coming year, and exposes the degree to which sustained consumer confidence will play a crucial cyclical role. 

Consumption demand in 2016 was supported by a sustained fall in the personal savings rate. Between Jan 2016 and December, the household savings rate had declined from 5.4% to 4.7% of total income, with annualized savings falling 8% yoy.  December’s 4.7% savings rate was as low as it has been since the start of 2014, and at the bottom end of the ‘new normal’ range. 

We must fear a reversion to the mean, with a rise in the savings rate depressing household consumption. But the dramatic impact of tax measures at end-2012/start-2013 makes it difficult to know what a reversion to the mean might look like. If one excludes both 2012 and 2013 from the data, then since 2010 the household savings rate has averaged 5.1% with a standard deviation of 0.3%.  If so, a reversion to the mean would involve a rise of approximately 40bps in the savings rate.  Now, a one percentage point rise in the savings rate would cut personal spending by approximately US$140bn, equivalent to 0.7% of GDP. On those calculations, a recovery of the savings rate would be expected to take approximately 30bps off GDP. 

In a ‘new normal’ world in which quarterly GDP yoy growth has averaged 2.1% with a standard deviation 0.7 percentage points, a cut of 30bps off GDP growth is material.  

Movements in consumer confidence are likely to be a major factor determining savings rates, and in particular, confidence in labour markets (ie, confidence in the security of income). This is just another reason why labour market indicators are the key to watch over the coming months. And among those, movements in labour participation will be the most important. I will look at that in Part III.