The Great Reversal Page 8
d Nokia is actually older than Finland. Nokia was founded in 1865. Finland only attained its independence in 1917 after over 700 years of Swedish rule and 109 years of Russian rule.
e For manufacturing, Feenstra and Weinstein (2017) construct such a measure. They use Census HHIs for the US and import data for foreign countries. We extend Feenstra and Weinstein’s data over time (until 2015) and outside the manufacturing sector. Outside manufacturing, neither Census nor foreign HHIs are available—so we have to use Compustat. We start with the “raw” HHIs from Compustat and adjust them to account for the domestic coverage of Compustat as well as the share of imports. The detailed calculations are described in Covarrubias, Gutiérrez, and Philippon (2019).
CHAPTER 4
The Decline of Investment and Productivity
IN CHAPTER 3, we demonstrated that since 2000, US industries have become more concentrated. Leaders’ market shares have become more persistent and their profit margins have increased. This pattern holds in all industries that are not heavily exposed to foreign competition, that is, most of the economy minus about half of the manufacturing sector.
We have argued that there are two leading explanations. One possible explanation is that industry leaders have become increasingly more efficient. This might explain why their market shares and their profits have grown. According to this view, which we called the Rise of Superstar Firms hypothesis, concentration is good news. The other explanation is that domestic competition has decreased and that leaders have become more entrenched. Their market shares are not threatened, and therefore they can charge high prices. According to this view, which we have called the Decreasing Domestic Competition hypothesis, concentration is bad news.
The two explanations are not mutually exclusive, in the sense that leaders can become more efficient and more entrenched at the same time. But they have opposite implications for efficiency, growth, welfare, and policy. The data that we have considered so far do not allow us to distinguish one explanation from the other. We must bring new data to bear. In this chapter, I will argue that the firms’ investment decisions offer clues about the driving forces behind rising concentration and profits.
Under the optimistic explanation, concentration is the consequence of increasingly efficient leading firms. To transmit this individual success to the broad economy, successful businesses should draw in more resources. They should hire and invest more. Do they?
Business Investment Has Been Low
Figure 4.1 shows that in recent years investment has been low relative to firms’ profits. Figure 4.1 shows the ratio of net investment (investment expenditures minus depreciation) to net operating surplus (gross surplus minus depreciation). Net investment is what matters for economic growth because it measures the change in capital from one year to the next.
There is a lot going on in Figure 4.1, so let us use the example from Chapter 3 to explain what these numbers mean. Recall that we imagined a firm with the following accounting information:
Assets
Revenues
Income
Depreciation
Taxes
Net investment
Dividends
$100
$150
$15
$5
$3
$2
$5
For this firm, we concluded that gross operating surplus (income) is $15. Depreciation is $5, so net operating surplus is $10. Gross investment is $7, and net investment is $2. Net investment over net operating surplus is 20 percent.
FIGURE 4.1 Net investment relative to net operating surplus
TABLE 4.1
Flow of Funds to Business Sector in 2014
Name
Value in 2014 ($ billions)
Corporate
(1)
Noncorporate
(2)
Business
(1 + 2)
Gross value added (PY)
$8,641
$3,147
$11,788
Stock of fixed capital (K)
$14,857
$6,126
$20,983
Consumption of fixed capital (CFK)
$1,286
$297
$1,583
Net operating surplus (PY−Wages−Tax−CFK)
$1,614
$1,697
$3,311
Gross fixed capital formation (I)
$1,610
$354
$1,964
Net fixed capital formation (I−CFK)
$325
$56
$381
Note: Stock of fixed capital is measured at replacement cost.
We can apply the same logic to the entire US economy. Table 4.1 summarizes the current account of the business sector in 2014. The starting point is gross value added, which means revenues minus the cost of intermediate inputs (materials) and energy (cost of electricity and so on). The gross value added of the US business sector in 2014 was $11.8 trillion, with a contribution of $8.6 trillion from incorporated businesses. Let’s call this gross value added PY: it is the product of the average price of goods and services (P) and the quantity of goods and services sold (Y), the letter used to denote real value added or real GDP in economic textbooks. To create this value added, the business sector uses a capital stock (K) of $21 trillion. This is measured at “replacement cost,” which means that it would cost $21 trillion to replace all the plants, warehouses, computers, vehicles, and equipment in the US business sector. Production wears out equipment, structures, and vehicles. Equipment and software can also become obsolete and be discarded. The sum of this wear and tear and obsolescence is called consumption of fixed capital (CFK), or more simply, depreciation. Replacing the depreciated capital at the end of the year costs $1.6 trillion. Finally, businesses spend money on employees’ wages and benefits. They also pay taxes on production. This leaves them with a net operating surplus of about $3.3 trillion. They use $381 billion to increase their capital stock. This figure is what we refer to as net investment.
In 2014, the ratio of net fixed capital formation ($381 billion) over net operating surplus ($3,311 billion) was 11.5 percent. As you can see from Figure 4.1, the average of this ratio between 1962 and 2001 is 20 percent. The average of the ratio from 2002 to 2015 is only 10 percent. In other words, over the first period, firms took 20 cents of each dollar of profits and plowed them back into their businesses by growing their capital stock.
In our simple example, with a net investment of 2 and starting from a capital stock of 100, the firm will have a capital stock of 102 next year. The growth rate of capital is 2 percent. This growth rate is important. When the capital stock grows, workers become more productive, and both labor demand and wages increase. In the long run, GDP and the capital stock tend to grow at the same rate.
In recent years firms have been plowing back into investment only a bit more than 10 cents for each dollar of profit. As a result, the growth of productive capital has been slow. Using the fixed asset tables from the Bureau of Economic Analysis (table 4.2 on the BEA website), we see that the growth rate of the capital stock of corporate businesses was 3.7 percent on average between 1962 and 2001, but only 1.9 percent on average between 2002 and 2012. You can see the trend decline in Figure 4.2.
How can we interpret this fact? Is this necessarily bad news? Perhaps firms are simply responding to market signals telling them the economy does not need more capital right now. Can we tell? Yes, we can, but first we need some theoretical background.
Why Do Firms Invest?
The goal of investment is to create (or replace) a long-lived valuable asset. This is in fact the exact definition of investment used in economic statistics.a Firms invest when they think that they need more long-lived assets. This can happen for two reasons: because firms perceive growing demand for their products, and because they want to innovate.
FIGURE 4.2 Declining growth of capital: growth rate of corporate businesse
s’ capital stock
When demand grows, firms usually start by increasing overtime: employees work longer hours, and the utilization rate of equipment increases. When the growth in demand is sustained, firms need to hire more capital and more labor.
Firms invest to expand their production capacity and satisfy a growing demand. Firms also invest to improve the range and quality of their products. In both cases, investment allows firms to increase their profits in the future. But what about the cost today? How can we compare uncertain future profits with current, known expenditures? This is where finance comes in.
Investment trades off future profits against current expenditures. The cost of financing the investment therefore plays a crucial role. Investment, by its very nature, is an intertemporal decision. You must decide how much to spend today in the hope of reaping uncertain benefits in the future. To assess the value of investment, you need to discount uncertain future payoffs. The financial markets allow you to do just that.
Consider the following example. You can invest in a project by buying an asset for $100 today. You think it will generate $12 in annual profits and depreciate by $6 each year. After the first year, for example, you pay $6 to replace the depreciated part, and you have a net income of $6. You do the same thing in the second year. The asset generates $6 year after year. The yield of this investment is 6 percent per year. Should you invest in this project? This depends on your funding cost. Imagine that you borrowed the $100 to buy the asset. If the interest rate on your loan is less than 6 percent, then it is a positive investment. Imagine that the interest rate on your loan is 5 percent. You repay $5 per year, and your net income after interest is $1 per year. At the rate of 5 percent, we say that the net present value (NPV) of this investment is positive. A perpetual income of $6 discounted at 5 percent is worth $6 / 0.05 = $120. The NPV is $120 − $100 = $20. The NPV is subject to the funding cost. If the funding cost was 7 percent instead of 5 percent, the NPV would be negative. A perpetual income of $6 discounted at 7 percent is worth $85.7. The NPV would be −$14.28, and you would not invest.
Making an investment decision therefore requires a complicated discounting of uncertain future cash flows and a comparison with the current cost of buying new capital. How can these decisions be made for the whole economy, with thousands of businesses, hundreds of billions of dollars of cash flows and investment costs? How can predictions be made not only one year ahead, but two, five, ten, or twenty?
It sounds like a daunting task, and it is. But if you think about it the right way and make some assumptions, you can see that accountants, together with stocks and bonds traders, have done it for us. This was the brilliant insight of Nobel Prize winner James Tobin. The measure that “looks at it the right way” is called Tobin’s q and is explained in Box 4.1.
Tobin’s insight is that the market value of the firm encodes all we want to know about the firm, at least as far as classical investment decisions are concerned. If accountants do their job properly, then we can measure the replacement cost of the firm’s fixed assets. If investors do their job properly, then we can measure the market value that the firm creates from these fixed assets. The difference between the market value and the replacement cost is the NPV of the firm. Tobin’s q is the ratio of the market value to the replacement cost. If q is more than one, the firm should scale up because each additional $1 of capital expenditure is worth q dollars. Tobin’s q contains a lot of information. In particular, Tobin’s q captures expectations about uncertain profits in the future as well as funding costs. The funding cost is directly reflected in the market value. If there is a crisis and investors freak out, you immediately see that funding costs rise, market values tank, and so does investment. If investors are optimistic, you see the opposite.
Box 4.1. Tobin’s q and the Fundamental Law of Investment
Let us return to our simple example of the firm with a stock of capital, measured at replacement cost, of $100. The firm pays out $5 this year and spends $2 on net investment. Imagine that the firm is financed only with equity: there are 100 shares, and investors value each share at $1.10, so the market value is $110. The shareholders own the capital stock as well as all future profits of the firm.
Tobin’s q is the ratio of the market value of the firm ($110) to the replacement cost of its capital ($100). For our firm, we therefore have q = 1.1. Tobin’s insight is that as long as q is more than one, the firm should invest. When q is more than one, the market values each additional $1 of capital at more than $1. By increasing its capital stock—by investing—the firm creates value.
How fast should the firm invest? Our firm could decide to increase its capital stock to $110 in one year, but that would require a lot of investment and disruption. Installing new equipment is costly and time consuming, so it makes sense to implement the investment plan over several years. In general, the annual rate of investment should follow this formula:
net investment rate = (q − 1) / investment time
This equation is the fundamental law of investment. The firm in our example uses an investment time parameter of five years. Since q − 1 = 0.1, with a five-year parameter, it chooses an investment rate of 0.1 / 5 = 2 percent. Our firm will grow slowly. It does not want to incur large adjustment costs.
It will invest 2 percent in the first year. Next year (assuming no news), its q will be 110 / 102, so it will invest 0.08 / 5 = 1.6 percent. The year after that, its q will be 110 / 103.6, and it will invest 1.2 percent. It will take a few more than four years to reach a book value of 105. It will never actually reach 110, but it will get closer each year.
We have assumed so far that the firm has only issued stocks, but Tobin’s insight also applies when the firm is financed with both stocks and bonds. In that case you need to add the value of the stocks and the value of the bonds to get the total market value of the firm. Imagine, then, that our firm has issued some bonds, and the value of the outstanding shares is $80 and the value of the outstanding bonds is $30. Bondholders and stockholders jointly own the capital and the future profits generated by the firm. The total value is still $80 + $30 = $110 and Tobin’s q is still 1.1.
How successful is Tobin’s q at explaining investment in practice? It depends on three main issues. First, you must measure capital and investment—in particular, intangible assets—correctly, using the methods of Peters and Taylor (2016), for instance. Second, q assumes that the market is rational, or at least that the managers, who make the investment decisions, and the investors, who price the shares, agree on the correct value for the firm. Bubbles in the market (internet startups in the late 1990s, crypto assets in recent years) disrupt this process. Third, and most relevant here, is the fact that q theory assumes that firms operate in competitive industries. If firms have market power, they will not want to invest until q = 1. They will restrict their expansion to maintain relatively high prices, and q will stay above 1. To put it differently, a decrease in competition would show up as an increasing gap in the relationship between q and investment. This is exactly what we see in the data.
The Growing Investment Gap
Let us put our theory to work. Figure 4.3 shows q and the investment rate. Tobin’s q has been rescaled to fit on the same graph. You can see that, as predicted by the fundamental law of investment, the two series are highly correlated: they move up and down together. But you can also see that after 2000, the investment rate seems to be lower than what one would predict based on q. In fact, if we cumulate the residual difference between the investment rate and q, we find that, by 2015, the capital stock is about 10 percent lower than it should be.b
FIGURE 4.3 Tobin’s q and investment. Tobin’s q is the market value of nonfinancial private businesses over the replacement cost of capital. Net investment is investment minus depreciation over the replacement cost of capital. Fitted values is investment predicted by q at the beginning of each year. Data source: BEA
This fact is interesting for us because this is exactly what the decrea
sing competition hypothesis would predict. The reason is intuitive. When q is above 1 in an industry, it means that there are rents left on the table. If the industry is competitive, these rents should be competed away: either incumbents would expand (as in our example), or new firms would enter. Over time, the capital stock would increase, and q would decrease toward 1. On the other hand, if the industry is not competitive, then investment would not increase as much, and q would remain above 1. If you believe that domestic competition has declined in the US economy, then you would expect a growing gap between q and investment, exactly as in Figure 4.3. Figure 4.3 supports the hypothesis that the US has experienced decreasing domestic competition.
We see a growing residual between Tobin’s q and net investment in the aggregate, but we can go a lot further. As we have discussed in previous chapters, concentration has increased more in some industries than in others. If the decreasing competition hypothesis is correct, then we would expect the investment-q residual to come from concentrating industries.
Figure 4.4 shows that this is exactly what we observe. We split industries into two groups based on the evolution of their HHIs. One group includes the ten industries where HHIs have increased the most; the other group the ten industries where HHIs have increased the least (as it turns out, HHIs are roughly constant in that group). We then estimate a fundamental law of investment for both groups of industries, and we compute the residuals. Figure 4.4 plots the cumulative residuals. You can see that the gap is essentially zero for nonconcentrating industries and more than 20 percent for the concentrating ones. On average, this is consistent with a 10 percent aggregate cumulative gap as argued earlier. The key point is that the aggregate gap comes entirely from concentrating industries.