When Investment Gets Real (and Netted)

"...These differences imply that a unit of real investment is now 18 percent cheaper relative to a unit of GDP than in 1970 and 20 percent cheaper relative to a unit of consumption goods. Within the category of investment, prices for non-construction investment goods have fallen quite dramatically relative to those for construction investment goods (Figure 10)....These differences imply that a unit of real investment is now 18 percent cheaper relative to a unit of GDP than in 1970 and 20 percent cheaper relative to a unit of consumption goods. Within the category of investment, prices for non-construction investment goods have fallen quite dramatically relative to those for construction investment goods (Figure 10)...."When investment gets real (and netted)Cardiff GarciaLearn moreFollow @cardiffgarcia

Expect plenty of debate on these matters for a long time to come. More inthe usual place.

Again, these are merely conjectures, and I didn’t even delve into the likely mismeasurements in the data or the role of policy throughout this period. Truthfully I’m not sure what’s going on and need to think about it more. Comments with alternative explanations would be appreciated.

Yet this doesn’t rule out the possibility that supply-side problems also have played a role. The IT improvements of the 1990s spread to other sectors as general-purpose technologies and contributed to widespread productivity acceleration. But once that process ran its course, productivity in these other sectors once again grew softly, as it has since the 1970s, even as these sectors’ relative share of the economy climbed:

Income inequality climbed and the labour share of income fell. The distributional outcomes were part of the problematic debt dynamics chronicled byAtif Mian and Amir Sufi, wherein broad-based growth instead was driven largely by middle- and lower-income households taking out debt against increasingly overvalued housing collateral. Then came the housing bust and the nightmarish cyclical collapse.

Hard to say. The impressive advancements in IT and heavily IT-using sectors were uniquely labour-saving, and they also were unpaired with the absorption of the displaced workers into good jobs in other sectors. (Globalisation alsoplayed a role.)

Given the continued deflation of IT goods, why didn’t the deepening resume? Why didn’t investment remain a larger share of the economyin real termsgiven that the goods were cheapening?

These are only guesses, but to start with the former, look again at Figure 13. There was a significant increase in the share of real, depreciation-adjusted, non-construction investment throughout the 1990s — representing significant capital deepening. But the progress reversed after the dot-com bubble burst:

Yet these investment trends could be consistent with elements ofboththe Secular Stagnation, which emphasizes a persistent demand shortfall, andthe supply-side productivity Great Stagnation idea.

The decline of the construction investment share, the inability of the economy to grow more strongly despite the technology-driven deflation of investment goods, and rising income inequality are all part of the Secular Stagnation thesis.

Ideally, companies would spend more of their cash either on new investment opportunities or on hiring people; expectations of strongernominaldemand growth in the future would help the chances of both. Such a development would also incentivise the riskier money managers to use less leverage, as unlevered return expectations would be higher.

Policymakers have spent the last six years seeking ways to disintermediate this relationship, both using regulatory tools and in some cases providing safer collateral directly. But a more fundamental macroeconomic fix is needed.

To explain the mechanism crudely, this pool contributed to the emergence of a financial system in which the repo desks at dealer banks acted as intermediaries between these cash pools seeking a safe money-like haven on one side; and aggressive, levered fixed-income investors (hedge funds, separate accounts, prop traders, absolute return bond funds, certain pension schemes, etc) on the other side. Among other uses, these investors channeled that leverage into securities financing.

In addition, the nominal trends do matter. The deflation in IT leaves big amounts of cash on the books of corporates. That cash is invested mainly for preservation. The steadily rising amount of this “leftover” money is one of the four new cash pools of the past two decades described in theexcellent paperby Zoltan Poszar thatwe coveredrecently. (The other three are the liquidity tranches of FX reserves managed by sovereigns, the cash managed by big institutional investors and asset managers, and the cash collateral reinvestment accounts of securities lenders.)

Is the secular decline in real construction investment as a share of GDP worrying? Is it just the natural consequence of an economy in which services increasingly are a bigger part? Maybe, though it also seems likely that part of this trend is being driven by changing demographics. As for the cyclical trend, the construction sector stillaccountsfor a large share of the remaining labour market slack.

All very interesting, though the right interpretation and policy conclusions remain unclear to me.

Cardiff here. The economists add that these trends are similar across the developed world.

All of this means that the data on real investment in the US look much healthier.In real terms, investment as a share of (real) GDP shows no sign of secular decline (Figure 11); and in particular, due to the price trends depicted in Figure 10, real non-construction investment shares appear to have secularly risen, doubling from roughly 6 percent in the early 1970s to above 12 percent in 2013 (Figure 12).

This is, in our view, an optimistic finding, considering that construction investment (especially residential) should, in theory, have less of an impact on potential growth than non-construction investment - due to the stronger technological propagation/spillover effects that typically come with investment in equipment and intellectual property.

Yet we should highlight that the real data shown in Figure 11 & Figure 12 measure investment in gross terms. Once depreciation is netted out, real investment trends look less optimistic, particularly for non-construction investment, where depreciation rates are the highest.

Indeed,real net non-construction investment as a share of (real) GDP doesn’t exhibit the rising long-term trend shown in Figure 12; but it also doesn’t appear to have secularly declined either (Figure 13).

These differences imply that a unit of real investment is now 18 percent cheaper relative to a unit of GDP than in 1970 and 20 percent cheaper relative to a unit of consumption goods. Within the category of investment, prices for non-construction investment goods have fallen quite dramatically relative to those for construction investment goods (Figure 10).

Figure 9 shows that the price deflator for investment goods has grown at a persistently slower rate than has the GDP deflator or the consumption deflator. From 1970-2013, the investment deflator grew by 3.3 percent per year on average, versus average yearly growth of 3.7 percent and 3.8 percent, respectively, for the GDP and PCE deflators.

From the note, emphasis ours, and click to enlarge the images:

So far, so unsurprising. More interesting is the outcome when the category-specific price deflators are applied.

They did find that the nominal decline of investment as a share of GDP was due to both construction and non-construction investment. But while the latter has been falling since the late 1990s, much of the recent decline in the share of nominal investment has been due to the former: in other words, to the housing bust.

In June, economists at Citi broke down the investment figures in the US and across advanced economies in three ways: real vs nominal, construction vs non-construction, and gross vs net of depreciation.

Most goods and services, of course, inflate rather than deflate over time, while deflation in IT allows companies to spend less of their money on this category without actually buying less of the IT goods. (A similar and related story can be told aboutdurable goods, whose component in the Personal Consumption Expenditure price index started deflating in the mid-1990s.)

A version of that question wasaskeda little while back by Matt Yglesias, who later built this chart (viaTyler Cowen) showing the decline in nominal IT investment in the United States as a share of nominal GDP:

Does the secular deflation of computers and related equipment skew the macroeconomic data on investment in the direction of irrelevance?

Cardiff writes mostly about US macroeconomic issues, with daily excursions into other topics about which he claim no expertise. Before Alphaville, Cardiff spent a little more than two years as a reporter at Dow Jones Financial News covering investment banking, asset management, and private equity. Along the way he has written freelance pieces on a variety of other topics from behavioural psychology to Muay Thai, the latter also being a personal interest that involves frequently getting kicked in the shins (and torso, and head).