The Great Reversal
Page 32
In matters of economic policy in particular, perfect can be the enemy of the good. In a globalized world, we must also acknowledge the value of listening to regulators in several jurisdictions. Dieselgate is one of Europe’s worst lobbying scandals. It exposed the deep corruption of European regulators by the car industry’s lobbyists. Without the efforts of the California Air Resources Board and the US Environmental Protection Agency, Europeans would have suffered from higher levels of pollution, and the cheaters would not have been punished. But think about it: it is not by chance that Dieselgate was uncovered in the US, and it is not by chance that the General Data Protection Regulation was implemented in Europe. In both cases, domestic politicians and regulators had been captured, but foreign regulators had not.
Principle 3: Protect Transparency, Privacy, and Data Ownership
You cannot think about competition in health, finance, transport, and many other industries without thinking about data, information, and privacy. Over and over, we have seen that the way oligopolies maintain high prices and avoid outrage and crackdowns is by hiding their fees. That is true of banks, credit card companies, pharmaceutical companies, hospitals, insurers, and internet platforms. You need to know what you pay, and why you pay it. And if you do not pay, you need to know which part of you is being sold.
As daunting as they may seem, all these problems have solutions. I strongly believe that the economic issues discussed in this book can be fixed. For this to happen, however, the world badly needs US policy makers to take the field. It is therefore disheartening to see the US conspicuously absent from the most important regulatory debate of the twenty-first century, that of privacy and data protection. The Chinese government is pursuing policies that violate individual privacy rights and bear a growing resemblance to George Orwell’s Big Brother. European policy makers are doing the best they can, but they are unlikely to succeed without active involvement from their US counterparts. To put it bluntly, solving the challenges of privacy and data protection requires a level of expertise that few institutions outside the US possess.
* * *
The economic challenges I have highlighted are universal. All countries need to face them. What sets the US apart is that it has more power and more responsibility. It has also lost its way more than other countries, perhaps because it started from a higher pedestal. This is not the first time a country has neglected what made it great in the first place. The Roman and Chinese empires lasted for centuries but fell apart when their leaders forgot the principles upon which the empires were built. The Siglo de Oro—the golden age of Spain—was brought down by nepotism, and by political and religious intolerance. The Dutch helped bring down the mighty Spanish empire and their republic became the foremost maritime and economic power of the seventeenth century, driven by creativity, entrepreneurship, and international openness. But they fell asleep at the wheel in trade, went to war with their neighbors, and allowed England to become the focal point of foreign investment and innovation. And the story goes on. Great powers rise, become complacent—or greedy—and fall.
The good news for the US is that this process, at least in historic terms, has only just begun, and the institutions that can arrest the fall remain in place. Yes, the US has neglected its free markets, but it has the opportunity to correct its mistakes. Yes, issues surrounding big data and privacy are difficult, but just as Europe has looked to the US in the past for ideas about improving competition, the US can now look to Europe to learn about protecting consumers’ privacy.
And finally, yes, there is too much money in politics and too much pressure from special interests clouding the judgment of lawmakers and regulators. This may be the most intractable problem we face, if only because it has the potential to scuttle solutions to bring back free markets. But there are many examples of advanced economies governed by legislators who don’t spend thirty hours every week dialing for dollars.
The US has overcome major challenges before. For more than a century, it has been at the leading edge of innovation, both private technological innovation and political and social innovation. US markets can and should regain their freedom.
* * *
a The novel The Leopard, made into a film in 1963, takes place in 1860, toward the end of the Risorgimento and the wars of Italian Unification. Garibaldi captures the Kingdom of the Two Sicilies. Garibaldi wants a republic but has to settle for a constitutional monarchy under Victor Emmanuel II, the first king of the united Italy.
b Trade Talks #66, December 2018. This podcast will convince you, whatever your political affiliations, that you can enjoy brilliant economic analysis in a nonpartisan way.
c Formally, I use a low labor-supply elasticity—the so-called Frisch elasticity—of 0.1. In theory, since profits and payouts (dividends) increase while wages decrease, some households will rationally decide not to work as hard as before. If the Frisch elasticity was high, labor supply would decrease significantly, and this would amplify my results: it would lead to larger losses in GDP, consumption, and labor income.
Appendix
A. INDUSTRY CLASSIFICATION: NAICS AND ISIC
If I ask you: Who are Sears’s competitors? You will quickly mention Walmart, Amazon, and a few others. But is that all? What are all the businesses that belong to the same industry? Economists and statisticians love to classify things. We classify firms into industries and industries into sectors. This appendix shows you what these names mean.
What is an industry? How do we classify firms across industries? We need to answer these basic questions if we want to study firms and industries. For instance, to compute the market share of a firm, we first need to figure out in what industry it operates (and then we need to worry about its physical location and such).
When we study the US, we will use the North American Industry Classification System (NAICS). It was developed jointly by the US, Canada, and Mexico in 1997 to replace the old Standard Industrial Classification (SIC) system. NAICS divides the economy into twenty sectors, and each sector is divided into industries. Table A.1 describes some important sectors of the US economy.
The goal of the classification is to group together economic units (factories, plants, stores) that have similar production processes. For instance, the NAICS information sector includes activities that transform and distribute information: broadcasting, publishing (books, newspapers, magazines), motion pictures, etc. The NAICS professional, scientific, and technical services sector covers activities where expertise (human capital) is the major input: lawyers, architects and interior designers, engineering services, advertising agencies, etc. The manufacturing sector is further divided into eighty-six industries at the four-digit level. Retail is divided into twenty-seven units and professional services into only nine.
Walmart is classified as retail. Its 2012-NAICS-3 code is 452 (general merchandise stores), and its NAICS-4 is 4529 (other general merchandise stores), which separates it from 4521 (department stores). Amazon is also in retail, but its codes are 454 (nonstore retailers) and 4541 (electronic shopping and mail-order houses.)
TABLE A.1.
NAICS Classification of Important Sectors of the US Economy
Selected Sector
Code
Definition
Example
Utilities
22
Generate, transmit & distribute gas, electricity, steam, water; sewage
22111 Electric power generation
Construction
23
Erect buildings & structures, repair & maintain
23731 Highway, street, and bridge construction
Manufacturing
31–33
Transform materials, substances, or components into new products
32541 Pharmaceutical and medicine manufacturing
Wholesale trade
42
Trade raw & intermediate materials, and goods for resale
42471 Petroleum bulk stations and terminals
Retail trade
44–45
Retail merchandise to the general public
44111 New car dealers
Transportation & warehousing
48–49
Transport passengers and cargo, store goods
481111 Scheduled passenger air transportation
Information
51
Distribute information and cultural products
51521 Cable
51721 Wireless carriers
Finance & insurance
51
Create and trade financial assets and insurance products
52311 Investment banking and securities dealing
Professional services
54
Provide scientific & technical services to organizations
54181 Advertising agencies
Health care & social assistance
62
Provide health care and social assistance to individuals
62121 Offices of dentists
Industry classification is difficult and imperfect. IBM started in computer manufacturing (334) but then moved toward professional services (541). In some databases, however, IBM’s codes have been the same since 1950. How does NAICS deal with the fact that many large companies operate in different industries? Essentially, it splits them up. NAICS classifies establishments, usually in a single physical location: factory, mill, store, hotel, movie theater, airline terminal, etc. An establishment is the smallest operating unit for which records of inputs (number of employees, wages, materials, capital) and output are available. The output may be sold or provided to the parent company. NAICS is a rather advanced system. For instance, a shop located in a hotel is a separate establishment classified in retail, while the hotel itself is in accommodation. In transportation or in telecommunications, the establishments are the permanent branch offices, terminals, and stations. A company often owns several establishments and therefore appears in more than one NAICS industry. NAICS includes specific rules to deal with vertical integration (i.e., steel mills that make steel but also produce steel castings) and joint production (a car dealership that both sells and repairs cars).
Finally, when we compare the US with other countries, we use the International Standard Industrial Classification system (ISIC), organized by the United Nations. The principles underlying ISIC are similar to the ones of NAICS, and the US, Canada, and Mexico have tried to create NAICS industries that do not cross ISIC two-digit boundaries. Some differences remain, however, especially when we use granular definitions.
B. UNDERSTANDING REAL GDP GROWTH
Nominal GDP for the US measures the market value of goods and services in the current year, using current prices. The problem with nominal GDP is that the base level of prices is arbitrary. US nominal GDP was $19.5 trillion in 2017. That is, if you measure it in dollars. If you measured it in pennies, it would be 1,950 trillion pennies. That’s a different number, but it obviously represents the same economic reality. We thus need to find a way to separate the concept of real GDP from that of GDP in some arbitrary unit of account.
Suppose that there are two goods, a and b. The quantity produced in year t are qa,t and qb,t, the prices are pa,t and pb,t. Nominal GDP is then:
Yt = pa,t qa,t + pb,t qb,t.
If the same goods and services were sold at the same prices every year, we could use nominal GDP to make meaningful comparisons. But everything changes: prices change, some goods appear, and some goods disappear. We discussed the issue of new goods earlier, in Chapter 2. The idea of real GDP is to net out the effect of prices. Historically, there have been two ways to do it.
Fixed-Weight Real GDP
The traditional way to define real GDP has been to fix a base year and use prices from that year, year 0. We can compute GDP in year t using prices from year 0:
Yt,0 = pa,0 qa,t + pb,0 qb,t.
Yt,0 is a measure of real GDP based on year 0. It measures what GDP would have been in year t if all the prices had remained the same as in year 0. It is what’s known as a Laspeyres index, meaning that it uses a fixed set of prices, or fixed weights.
Yt,0 / Y0 is a measure of real growth between year 0 and year t. This approach was used in the US until 1996. It is rather simple to explain, but it has one big drawback: the number you get for “real growth” depends on the arbitrary choice of the base year. For instance, the growth rate of the US economy in 1998 was 4.5 percent using 1995 as the base year, but 6.5 percent using 1990 prices, 18.8 percent using 1980 prices, and 37.4 percent using 1970 prices (Whelan, 2000). This problem is known as substitution bias, and it happens when there are large changes in relative prices over time.
Chained Indexes
The problem of using a base year is that its prices become obsolete. So why don’t we use last year’s prices instead? That is the basic idea of chained indexes:
Yt,t − 1 = pa,t − 1 qa,t + pb,t − 1 qb,t.
This is the GDP this year using last year’s prices, and we can compute the growth rate in year t as:
The growth rate is the Laspeyres growth with the previous year as base year. This is almost how growth is computed today. It is slightly more advanced because, if you think about it, when you compute growth between t − 1 and t, you might want to treat t − 1 and t in a more symmetric way. You can compute GDP at t − 1 using t prices as Yt − 1,t = pa,t qa,t − 1 + pb,t qb,t − 1. Then you can compute growth as:
This is called a Paasche index. Both growth estimates are sensible, so why not use the geometric average? That is the method employed by the US Commerce Department’s Bureau of Economic Analysis to produce the Fisher index of real growth:
This is the headline number you hear about each quarter when the BEA releases its estimate of the growth rate of GDP.
C. REAL EXCHANGE RATES AND BALASSA-SAMUELSON
There are two types of exchange rates: a financial rate based on the foreign exchange (FOREX) market and a PPP rate based on local prices (International Comparison Program [ICP] PPP, or Big Mac PPP). Before we can compare prices around the world, we need to pause and think about how exchange rates are determined. The theory of purchasing power parity (PPP) says that, in the long run, the exchange rate adjusts in such a way that the law of one price (LOOP) holds for a basket of goods (see Chapter 7). What is the link between PPP and LOOP? The LOOP applies to individual goods (say, a pair of shoes), and PPP applies to the general price index (the price of a basket of goods). Clearly, if the LOOP holds for each good, then PPP will also hold for the basket of goods. Some prices might be off individually, but PPP might still work reasonably well for the average basket. So let’s imagine a representative basket of goods sold in both Europe and the US. That would include food, cars, electronics, etc. If PUS is the cost of that basket in dollars in the US and PEU is the cost of that same basket in euros in Europe, then the theory of PPP says that the euro / dollar exchange rate should gravitate toward the PPP rate:
EPPP = PUS / PEU.
Is there support for the PPP theory of exchange rates? It’s the classic frustrating story of the half-empty glass. In the short term, exchange rates move for many apparently random reasons, largely unrelated to relative prices. And local prices are slow to adjust to changes in nominal exchange rates. PPP can only be a theory of exchange rates in the long run. Even then, the support is rather weak in the sense that the prices of similar baskets can remain substantially different for a long time.
FIGURE A.1 Nominal and real exchange rates. The real exchange rate (RER) is the ratio of the nominal rate to the PPP rate. When the RER rate is less than one, the euro is cheap. According to this view, the euro was somewhat expensive in 2007–2008, but has been cheap since 2015. Volatility is the sample standard deviation of the series.
Real Exchange Rates
There is support, however, for predicting the evolution of exchange rates: if a country is expensive, its currency tends to depreciate. To quantify deviations from PPP, we define the euro real exchange rate (RER) as
:
RER = EMARKET / EPPP.
One way to state the PPP theory of exchange rates is that the real exchange rate should tend to 1 in the long run. Figure A.1 shows the nominal and real exchange rates of the euro against the dollar, i.e., the comparison between the US and EA19. EPPP in 2017 is $1.33, and the market rate is $1.13. The RER is therefore 0.84. The theory predicts that the RER should increase—that is, the euro should appreciate over the next few years (or, equivalently, that the dollar should depreciate).
Notice that the RER is not much less volatile than the nominal exchange rate. The volatility of the market rate is $0.17, and it is 0.12 for the ICP-based RER. In the short term, movement in relative prices does not cancel much of the movements in financial rates.
The Balassa-Samuelson Effect
Wages are higher in rich countries, so local prices are higher. This circumstance is called the Balassa-Samuelson effect. Many goods are not traded internationally (haircuts, for example), and even the ones that are traded incur local distribution costs (labor costs and commercial real estate rents). The original figure from Balassa’s 1964 paper shows that rich countries tend to have more expensive exchange rates than poor countries. Their prices are higher than one would predict based on the level of the market exchange rates. This is consistent with the haircut example that we used at the beginning of Chapter 7.
Angus Deaton and Alan Heston (2010) look at the same data forty-five years later and confirm the Balassa-Samuelson effect. In most emerging markets, the ratio of the market and PPP US dollar exchange rates is between 2 and 4, because nontraded goods and services are cheap in low-income countries even though the prices of tradable goods (machines, etc.) are broadly similar. These differences mean that developing countries are richer on a PPP basis than on a market-rate basis. Incidentally, it means that the world appears somewhat less unequal when we view it from a PPP perspective.