Every month the ISM releases two reports-based surveys of purchasing and supply executives in both the manufacturing and non-manufacturing sectors of the economy. Every month, I try to provide another angle to view the reports. This month’s perspective is provided by Mike Shedlock (Mish) with his post entitled “Manufacturing ISM Weaker Than it Looks; Digging Into the Numbers; Inventory Restocking Accounts for Much of the Rise.” The part that is most important in the discussion here is his conclusion under “Addendum – Reply from ISM.”
Since it is equal weighted of five components [New Orders, Production, Employment, Supplier Deliveries, and Inventories], the effect of inventories is 5.4 divide by five, or 1.08 (1.1) of the overall 1.8 rise as noted by Goldman Sachs.
Mish is correct that over 60% (technically 1.08/1.76= 61.4%) of the rise is attributed to the predominant increase in the inventory index. But the division by 5 is only correct if all the numbers are of the same sign. And in this case, they all increased, meaning faster rates of growth for each of the factors. Below is the exact formula and index numbers to calculate the difference between consecutive months:
ΔPMI = 0.2*ΔNewOrd + 0.2*ΔProd + 0.2*ΔEmp + 0.2*ΔSupDel + 0.2*ΔInv
1.76 = 0.2*0.6+0.2*0.5+0.2*1.7+0.2*0.6+0.2*5.4
The fact that they all increased shows that reversal is broad across all indexes although not significant in most. As compared to May, this is very welcome news as all components of the PMI were down. Below is the same equation with the May numbers.
-6.88 = 0.2*-10.7+0.2*-9.8+0.2*-4.5+0.2*-4.5+0.2*-4.9 (Reported as -6.9%)
The Macro View has dealt more with the indexes of Production, Employment, and New Orders. They represent the heart of the issue of economic growth with production; the continuing jobs recovery going forward in employment; and a forward look at production and growth of the economy with new orders. I have also been concerned about the price indexes, as this could signify structural rigidity and portent inflation, and the exports indexes, as this could be an exogenous stimulus to the economy.
One problem with using the inventory indexes is that they do not differentiate between planned investments and unplanned investments through inventory changes. A planned increase in inventory levels can signify greater real economic investment rather than a potential slowdown in aggregate consumption. All macroeconomic textbooks I have read break down investments into the subcategory of inventory investments. The following equation and discussion is derived from "Macroeconomics: Theories and Policies," second edition by Richard Froyen.
ΔINV(t) = λ(γSexp-INV(t-1)) + λ(Sexp-Sact)
The equation states that changes in the inventory levels are derived from the expected sales level (Sexp) times a proportional factor (γ) plus the difference between expected and actual sales (Sact). The proportional factor is to adjust inventories to the desired level in a gradual process. Since sales in the aggregate is a function of aggregate income, then inventory levels tend to be procyclical to general economic cycles with a lag. As aggregate income increases then general sales levels increase and businesses then increase inventory investments; conversely as incomes decrease and sales follow, businesses tend to decrease inventory investments.
Putting on the Neo-Keynesian hat would point out that a central government can possibly counter this effect on the economy. Fiscally, it can increase its own investment levels and increase aggregate spending (sales) in the short run. Monetary policy can also counter the effects of the business inventory cycle. Quoting from Macroeconomics:
Higher interest rates would increase the carrying cost of inventories and therefore reduce inventory investment. Monetary policy, by changing the interest rate, could then potentially eliminate the cyclical volatility of inventory investment and thereby lessen the overall cyclical variation in GNP [GDP].
Volatility of the Inventory Index
Since we cannot differentiate between planned and unplanned inventory investments, the volatility of the index becomes even more important in determining how much attention we pay to any individual monthly report. Below is a graph showing the changes in the inventory index from the previous month, the percentage differences from respondents reporting higher inventories than those reporting lower inventories (Differences in net), and the PMI index, minus 50 to scale it to the other numbers and show the differences between the break-even point (zero).
(Click to enlarge)
This graph shows that monthly differences in the inventory index appear to be a stochastic variable with no discernible pattern or trend. Inventory levels can be a lagging indicator as restocking occurs. Mish agrees with ZeroHedge on this theory as stated, “If anything, an increase in inventories is a negative for future activity.” Or it can be a leading indicator, as the general business climate has improved marked by a small but significant increase in 10 year bond rates as noted by Paul Krugman in "Interest Rates: A Dowist Perspective."
Even if it was restocking last month, the graph shows that “restocking” is occurring regularly. From February to May of this year, the PMI has been declining but inventory has swung wildly up and down as much as 5 percentage points in each direction.
Overall, the ISM manufacturing report is robust as the 5 indexes that make up the PMI reversed direction and increased at a faster rate of growth. May’s report was the downer and hopefully June’s report signifies that was just a slow patch in the economic recovery. It would have been better if the new orders index was stronger than a weak 51.6% and production was above the mid-50s range, but growth is growth at anything above 50%. Thus we should not lament the gyrations of the inventory index but look at the total picture of the reports.
Mish could have also found out the information about weighting the indexes from the ISM website at Frequently Asked Questions. The two breakdowns for manufacturing and non-manufacturing indexes are:
Q: How is the PMI calculated, and what does it mean?
A: The PMI is a composite index based on the seasonally adjusted diffusion indexes for the following five indicators at equal weights:
New Orders 20%
Supplier Deliveries 20%
Q: Which index in the Non-Manufacturing Report On Business® is a composite index or equivalent to the PMI?
A: Beginning in January 2008, ISM began calculating a composite index for the Non-Manufacturing sector. The new Non-Manufacturing Index, NMI, consists of:
Business Activity 25%
New Orders 25%
Supplier Deliveries 25%
Note that inventory is the one index not included in the NMI. The obvious answer is that services industries hold less inventory, but I suspect that it has more to do with its reliability as an indicator of business activity and the subsequent lack of being statistically significant at least for the services industries.