Slightly disappointing as it came with rave reviews from a colleague; my expectations were probably too hyped. Could've been revolutionary in its time -- published a massive 4 years ago, but the market adopts new practices like 'snap' -- but I think a lot of what he discusses are already integrated into leading i-bank analytical work so is perhaps less stunning than what I expected. Still, a logical and clear read, with some interesting points, below:
1). Basic schematic of US economy (actual figure more complicated than I can depict here):
Real consumer spending => w 0-6m lag
Industrial Prod and Services => w 6-12m lag
Real Capital spending
All 3 above => Corporate profits => Stock market and Employment
IP always swings more than Real Con, b/c of inventory effects that increase amplitude of swing. Even more so for capital goods as they are further up the value chain. The volatility of production (at consumer, inventory supplier, and capital goods supplier levels) all lead to big swings in corporate profits at respective levels, leading to stock price changes. Declines in Real Con don't always lead to recession, but do almost always lead to bearish stock markets. (Note he describes the business/stock cycle here, not structural changes -- I think this is a big point that could cause one to misinterpret the strength and applicability of Ellis conclusions).
2). Track
(i) YoY rates, not actual levels, not MoM or QoQ.
(ii) use 2 charts, showing cause and effect
Other tips: using a rolling 3m avg to smooth monthly data, use different scales, have vertical and horizontal lines to aid perception
3). Unit purchasing power, or real wages, determines consumption of 93-96% of the workforce (assuming 3-7% unemployment; I adjust this too for those out of workforce). Hence, changes in real wages are more significant factor in driving consumption, than are changes in unemployment.
If you consider both
(i) YoY changes in Real Wages (= Nominal hourly wages x PCE deflator which is CPI), as well as
(ii) YoY changes in employment,
this gives you most of what is driving US real consumption. The third factor is consumer credit, discussed in the appendix.
4). The role of credit -- interestingly, when he charts real wage change against real consumption change (graph 10-10, p 134), it looks to me that real consumption change always stays above real wage change. I.e. in good times, it improves faster and reaches higher levels of change, than in bad times, when the lowest rate to which it falls is the rate of change of real wages. To me, this shows that structurally, the ratio of real consumption is rising relative to real wages, which reflects the role of credit in the US economy.
Ellis comes to this conclusion too -- that credit amplifies the impact of real wage growth. Interestingly, he presents a simply numerical eg that shows persuasively that it is new borrowing (ie. change in debt) that adds to consumer purchasing power, and thus it is the rate of change of new borrowing -- ie, the second derivative of credit levels -- that an analyst should track.
He also says that borrowing increases strongly when employment growth is strongest and consumers are confident on outlook to take on more debt => this means though that credit growth is a coincident to lagging indicator, not a leading one.
But he doesn't discuss -- and no fault of his, since he couldn't have foreseen the GFC in 2009 -- structural changes in the level of credit; e.g. qns like when the interest cost burden of debt cripples an economy, or what happens in a structural deleveraging cycle. (To his credit, he does not in his final chapters that US discount rate, prime rate, and 10 yr bond yields were, in 2004, at historic lows, and cautions investors to watch for higher yields and a structural bear market.) Possibly, this omission could throw off his conclusions -- real wage changes would still lead real consumption, but what changes is that the contribution of credit is negative, rather than positive, and while real consumption still follows the cycle, on a through the cycle basis it trends downward, rather than upward.
5). Interest rates -- also a good lead indicator, from his charts. But he argues that only about 7% of consumption in the US is tied to interest rates -- big consumer durables, home furnishings, jewelry and the like. However, he says, it is because interest rates are driven by, and lag slightly, inflation, which does affect real wages.
He also notes that interest rates drive expected returns on different asset classes, but the book is less concerned about strategic asset allocation and he doesn't delve into this.