Tariffs have far-reaching effects that strike some as counter-intuitive, but they are real forces nevertheless. Much like any selective excise tax, tariffs reduce the quantity demanded of the taxed good; buyers (importers) pay more, but sellers of the good (foreign exporters) extract less revenue. Suppose those sellers happen to be the primary buyers of what you produce. Because they have less to spend, you also will earn less revenue.
The Lerner Effect
The imposition of tariffs by the U.S. means that foreigners have fewer dollars to spend on exports from the U.S. (as well as fewer dollars to invest in the U.S. assets like Treasury bonds, stocks, and physical capital). That much is true without any change in the exchange rate. However, lower imports also imply a stronger dollar, further eroding the ability of foreigners to purchase U.S. exports.
The implications of the import tariff for U.S. exports may be even more starkly negative. Scott Sumner discusses an economic principle called Lerner Symmetry: a tax on imports can be the exact equivalent of a tax on exports! That’s because two-directional trade flows rely on two-directional flows of income.
Note that this has nothing to do with foreign retaliation against U.S. trade policy, although that will also hurt U.S. exporters. Nor is it a consequence of the very real cost increase that tariffs impose on U.S. export manufacturers who require foreign inputs. That’s a separate issue. Lerner Symmetry is simply part of the mechanics of trade flows in response to a one-sided tariff shock.
Assumptions For Lerner Symmetry
Scott Sumner enumerated certain conditions that must be in place for full Lerner Symmetry. While they might seem strict, the Lerner effect is nevertheless powerful under relaxed assumptions (though somewhat weaker than full Lerner Symmetry).
As Sumner puts it, while full Lerner Symmetry requires perfect competition, nearly all markets are “workably competitive”. In the longer-run, assumptions of price flexibility and full employment are anything but outlandish. Complete non-retaliation is an unrealistic assumption, given the breadth and scale of the Trump tariffs. Some countries will retaliate, but not all, and it is certainly not in their best interests to do so. The assumption of balanced trade is one and the same as the assumption of no capital flows; a departure from these “two” assumptions weakens the symmetry between tariffs and export taxes because a reduction in capital flows takes up some of the slack from lower revenue earned by foreign producers.
Trump Tariff Impacts
So here we are, after large hikes in tariffs and perhaps more on the way. Or perhaps more exceptions will be carved out for favored supplicants in return for concessions of one kind or another. All that is economically and ethically foul.
But how are imports and exports faring? Here I’ll quote the Yale Budget Lab’s (YBL) September 26th report on tariffs, which includes the chart shown at the top of this post:
“Consumers face an overall average effective tariff rate of 17.9%, the highest since 1934. After consumption shifts, the average tariff rate will be 16.7%, the highest since 1936. …
The post-substitution price increase settles at 1.4%, a $1,900 loss per household.“
The “post-substitution” modifier refers to the fact that price increases caused by tariffs would be somewhat larger but for consumers’ attempts to find lower-priced domestic substitutes. Suppose the PCE deflator ends 2025 with a 2.8% annual increase. The YBL’s price estimate implies that absent the Trump tariffs, the PCE would have increased 1.4%. If that seems small to you (and the tariff effect seems large to you), recall that monetary policy has been and remains moderately restrictive, so we might have expected some tapering in the PCE without tariffs.
We also know that the early effects of the tariffs have been dominated by thinner margins earned by businesses on imported goods. Those firms have been swallowing a large portion of the tariff burden, but they will increasingly attempt to pass the added costs into prices.
But back to the main topic … what about exports? Unfortunately, the data is subject to lags and revisions, so it’s too early to say much. However, we know exports won’t decline as much as imports, given the lack of complete Lerner symmetry. YBL predicts a drop in exports of 14%, but that includes retaliatory effects. In August the WTO predicted only about a 4% decline, which would be about half the decline in imports.
Seeking Compensatory Rents
More telling perhaps, and it may or may not be a better indicator of the Lerner effect, is the clamoring for relief by American farmers who face diminished export opportunities. As Tyler Cowen says, “Lerner Symmetry Bites”. Other industries will feel the pinch, but many are likely preoccupied with the more immediate problem of increases in the direct cost of imported materials and components.
The farm lobby is certainly on its toes. The Trump Administration is now asking U.S. taxpayers to subsidize soybean producers to the tune of $15 billion. Those exporting farmers are undoubtedly victimized by tariffs. But so much for deficit reduction! More from Cowen:
“Using tariff revenue to subsidize the losses of exporters is a textbook illustration of Lerner Symmetry because the export losses flow directly from the tax on imports! The irony is that President Trump parades the subsidies as a victory while in fact they are simply damage control for a policy he created.“
A List of Harms
Tariffs are as distortionary as any other selective excise tax. They restrict choice and penalize domestic consumers and businesses, whose judgement of cost and quality happen to favor goods from abroad. Tariffs create cost and price pressures in some industries that both erode profit margins and reduce real incomes. For consumers, a tariff is a regressive tax, harming the poor disproportionately.
Tariffs also diminish foreign flows of capital to the U.S., slowing the long-term growth of the economy as well as productivity growth and real wages. And the Lerner effect implies that tariffs harm U.S. exporters by reducing the dollars available to foreigners for purchasing goods from the U.S. In these several ways, Americans are made worse off by tariffs.
We now see attempts to cover for the damage done by tariffs by subsidizing the victims. A “tariff dividend” to consumers? Subsidies to exporters harmed by the Lerner effect? In both cases, we would forego the opportunity to pay down the bloated public debt. Thus, the American taxpayer will be penalized as well.
There’s a hopeful narrative making the rounds that artificial intelligence will prove to be such a boon to the economy that we need not worry about high levels of government debt. AI investment is already having a substantial economic impact. Jason Thomas of Carlyle says that AI capital expenditures on such things as data centers, hardware, and supporting infrastructure account for about a third of second quarter GDP growth (preliminarily a 3% annual rate). Furthermore, he says relevant orders are growing at an annual rate of about 40%. The capex boom may continue for a number of years before leveling off. In the meantime, we’ll begin to see whether AI is capable of boosting productivity more broadly.
Unfortunately, even with this kind of investment stimulus, there’s no assurance that AI will create adequate economic growth and tax revenue to end federal deficits, let alone pay down the $37 trillion public debt. That thinking puts too much faith in a technology that is unproven as a long-term economic engine. It would also be a naive attitude toward managing debt that now carries an annual interest cost of almost $1 trillion, accounting for about half of the federal budget deficit.
Boom Times?
Predictions of AI’s long-term macro impact are all over the map. Goldman Sachs estimates a boost in global GDP of 7% over 10 years, which is not exactly aggressive. Daren Acemoglu has beenevenmore conservative, estimating a gain of 0.7% in total factor productivity over 10 years. Tyler Cowen has been skeptical about the impact of AI on economic growth. For an even more pessimistic take see these comments.
In July, however, Seth Benzell of the Stanford Digital Economy Lab discussed some simulations showing impressive AI-induced growth (see chart at top). The simulations project additional U.S. GDP growth of between 1% – 3% annually over the next 75 years! The largest boost in growth occurs now through the 2050s. This would produce a major advance in living standards. It would also eliminate the federal deficit and cure our massive entitlement insolvency, but the result comes with heavy qualifications. In fact, Benzell ultimately throws cold water on the notion that AI growth will be strong enough to reduce or even stabilize the public debt to GDP ratio.
The Scarcity Spoiler
The big hitch has to do with the scarcity of capital, which I’ve described asanimpediment to widespread AI application. Competition for capital will drive interest rates up (3% – 4%, according to Benzell’s model). Ongoing needs for federal financing intensify that effect. But it might not be so bad, according to Benzell, if climbing rates are accompanied by heightened productivity powered by AI. Then, tax receipts just might keep-up with or exceed the explosion in the government’s interest obligations.
A further complication cited by Benzell lurks in insatiable demands for public spending, and politicians who simply can’t resist the temptation to buy votes via public largesse. Indeed, as we’ve already seen, government will try to get in on the AI action, channeling taxpayer funds into projects deemed to be in the public interest. And if there are segments of the work force whose jobs are eliminated by AI, there will be pressure for public support. So even if AI succeeds in generating large gains in productivity and tax revenue, there’s very little chance we’ll see a contagion of fiscal discipline in Washington DC. This will put more upward pressure on interest rates, giving rise to the typical crowding out phenomenon, curtailing private investment in AI.
Playing Catch-Up
The capex boom must precede much of the hoped-for growth in productivity from AI. Financing comes first, which means that rates are likely to rise sooner than productivity gains can be expected. And again, competition from government borrowing will crowd out some private AI investment, slowing potential AI-induced increases in tax revenue.
There’s no chance of the converse: that AI investment will crowd out government borrowing! That kind of responsiveness is not what we typically see from politicians. It’s more likely that ballooning interest costs and deficits generally will provoke even more undesirable policy moves, such as money printing or rate ceilings.
The upshot is that higher interest rates will cause deficits to balloon before tax receipts can catch up. And as for tax receipts, the intangibility of AI will create opportunities for tax flight to more favorable jurisdictions, a point well understood by Benzell. As attorneys Bradford S. Cohen and Megan Jones put it:
“Digital assets can be harder to find and more easily shifted offshore, limiting the tax reach of the U.S. government.”
AI Growth Realism
Benzell’s trepidation about our future fiscal imbalances is well founded. However, I also think Benzell’s modeled results, which represent a starting point in his analysis of AI and the public debt, are too optimistic an assessment of AI’s potential to boost growth. As he says himself,
“… many of the benefits from AI may come in the form of intangible improvements in digital consumption goods. … This might be real growth, that really raises welfare, but will be hard to tax or even measure.”
This is unlikely to register as an enhancement to productivity. Yet Benzell somehow buys into the argument that AI will lead to high levels of unemployment. That’s one of his reasons for expecting higher deficits.
My view is that AI will displace workers in some occupations, but it is unlikely to put large numbers of humans permanently out of work and into state support. That’s because the opportunity cost of many AI applications is and will remain quite high. It will have to compete for financing not only with government and more traditional capex projects, but with various forms of itself. This will limit both the growth we are likely to reap from AI and losses of human jobs.
Sovereign Wealth Fund
I have one other bone to pick with Benzell’s post. That’s in regard to his eagerness to see the government create a sovereign wealth fund. Here is his concluding paragraph:
“Instead of contemplating a larger debt, we should instead be talking about a national sovereign wealth fund, that could ‘own the robots on behalf of the people’. This would both boost output and welfare, and put the welfare system on an indefinitely sustainable path.”
Whether the government sells federal assets or collects booty from other kinds of “deals”, the very idea of accumulating risk assets in a sovereign wealth fund undermines the objective to reduce debt. It will be a struggle for a sovereign wealth fund to consistently earn cash returns to compensate for interest costs and pay down the debt. This is especially unwise given the risk of rising rates. Furthermore, government interests in otherwise private concerns will bring cronyism, displacement of market forces by central planning, and a politicization of economic affairs. Just pay off the debt with whatever receipts become available. This will free up savings for investment in AI capital and hasten the hoped-for boom in productivity.
Summary
AI’s contribution to economic growth probably will be inadequate and come too late to end government budget deficits and reduce our burgeoning public debt. To think otherwise seems far fetched in light of our historical inability to restrain the growth of federal spending. Interest on the federal debt already accounts for about half of the annual budget deficit. Refinancing the existing public debt will entail much higher costs if AI capex continues to grow aggressively, pushing interest rates higher. These dynamics make it pretty clear that AI won’t provide an easy fix for federal deficits and debt. In fact, ongoing federal borrowing needs will sop up savings needed for AI development and diffusion, even as the capital needed for AI drives up the cost of funds to the government. It’s a shame that AI won’t be able to crowd out government.
As a long-time user of macroeconomic statistics, I admit to longstanding doubts about their accuracy and usefulness for policymaking. Almost any economist would admit to the former, not to mention the many well known conceptual shortcomings in government economic statistics. However, few dare question the use of most macro aggregates in the modeling and discussion of policy actions. One might think conceptual soundness and a reasonable degree of accuracy would be requirements for serious policy deliberation, but uncertainties are almost exclusively couched in terms of future macro developments; they seldom address variances around measures of the present state of affairs. In many respects, we don’t even know where we are, let alone where we’re going!
Early and Latter Day Admonitions
In the first of a pair of articles, Reuven Brenner discusses the hazards of basing policy decisions on economic aggregates, including critiques of these statistics by a few esteemed economists of the past. The most celebrated developer of national income accounting, Simon Kuznets, was clear in expressing his reservations about the continuity of the U.S. National Income and Product Accounts during the transition to a peacetime economy after World War II. The government controlled a large share of economic activity and prices during the war, largely suspending the market mechanism. After the war, market pricing and private decision-making quickly replaced government and military planners. Thus, the national accounts began to reflect values of production inherent in market prices. That didn’t necessarily imply accuracy, however, as the accounts relied (and still do) on survey information and a raft of assumptions.
The point is that the post-war economic results were not remotely comparable to the data from a wartime economy. Comparisons and growth rates over this span are essentially meaningless. As Brenner notes, the same can be said of the period during and after the pandemic in 2020-21. Activity in many sectors completely shut down. In many cases prices were simply not calculable, and yet the government published aggregates throughout as if everything was business as usual.
More than a decade after Kuznets, the game theorists Oskar Morgenstern and John von Neumann both argued that the calculations of economic aggregates are subject to huge degrees of error. They insisted that the government should never publish such data without also providing broad error bands.
Morgenstern delineated several reasons for the inaccuracies inherent in aggregate economic data. These include sampling errors, both private and political incentives to misreport, systematic biases introduced by interview processes, and inherent difficulties in classifying components of production. Also, myriad assumptions must be fed into the calculation of most economic aggregates. A classic example is the thorny imputation of services provided by owner-occupied homes (akin to the value of services generated by rental units to their occupants). More recently. Charles Manski reemphasized Morganstern’s concerns about the aggregates, reaching similar conclusions as to the wisdom of publishing wide ranges of uncertainty.
Real or Unreal?
Estimates of real spending and production are subject to even larger errors than estimates of nominal values. The latter are far simpler to measure, to the extent that they represent a simple adding up of current amounts spent (or income earned) over the course of a given time period. In other words, nominal aggregates represent the sum of prices times quantities. To estimate real quantities, nominal values must be adjusted (deflated) by price aggregates, the measurement of which are fraught with difficulties. Spending patterns change dramatically over time as preferences shift; technology advances, new goods and services replace others, and the qualities of goods and services evolve. A “unit of output” today is usually far different than what it was in the past, and adjusting prices for those changes is a notorious challenge.
This difficulty offers a strong rationale for relying on nominal quantities, rather than real quantities, in crafting certain kinds of policy. Perhaps the best example of the former is so-calledmarket monetarism and monetary policy guided by nominal GDP-level targeting, as championed by Scott Sumner.
Government’s Contribution
Another fundamental qualm is the inconsistency between data on government’s contribution to aggregate production versus private sector contributions. This is similar in spirit to Kuznets’ original critique. Private spending is valued at market prices of final output, whereas government spending is often valued at administered prices or at input cost.
An even deeper objection is that much of the value of government output is already subsumed in the value of private production. Kuznets himself thought so! For example, to choose two examples, public infrastructure and law enforcement contribute services which enhance the private sector’s ability to reliably produce and deliver goods to market. To add the government’s “output” of these services separately to the aggregate value of private production is to double count in a very real sense. Even Tyler Cowen is willing to entertain the notion that including defense spending in GDP is double counting. The article to which he links goes further than that.
Nevertheless, our aggregate measures allow for government spending to drive fluctuations in our estimates of GDP growth from one period to another. It’s reasonable to argue that government spending should be reported as a separate measure from private GDP.
But what about the well known Keynesian assertion that an increase in government spending will lift output by some multiple of the change? That proposition is considered valid (by Keynesians) only when resources are idle. Of course, today we see steady growth of government even at full employment, so the government’s effort to commandeer resources creates scarcity that crowds out private activity.
Measurement and Policy Uncertainty
Acting on published estimates of economic aggregates is hazardous for a number of other reasons. Perhaps the most basic is that these aggregates are backward-looking. A policy activist would surely agree that interventions should be crafted in recognition of concurrent data (were it available) or, even better, on the basis of reliable predictions of the future. Financial market prices are probably the best source of such forward-looking information.
In addition, revising the estimates of aggregates and their underlying data is an ongoing process. Initial published estimates are almost always based on incomplete data. Then the estimates can change substantially over subsequent months, underscoring uncertainty about the state of the economy. It is not uncommon to witness consistent biases over time in initial estimates, further undermining the credibility of the effort.
Even worse, substantial annual revisions and so-called “benchmark revisions” are made to aggregates like GDP, inflation, and employment data. Sometimes these revisions alter economic history substantially, such as the occurrence and timing of recessions. All this implies that decisions made on the basis of initial or interim estimates are potentially counterproductive (and on a long enough timeline, every aggregate is an “interim” estimate). At a minimum, the variable nature of revisions, which is an unavoidable aspect of publishing aggregate statistics, magnifies policy uncertainty.
Case Studies?
Brenner cites two historical episodes as support for his argument that aggregates are best ignored by policymakers. They are interesting anecdotes, but he gives few details and they hardly constitute proof of his thesis. In 1961, Hong Kong’s financial secretary stopped publishing all but “the most rudimentary statistics”. Combined with essentially non-interventionist policy including low tax rates, Hong Kong ran off three decades of impressive growth. On the other hand, Argentina’s long economic slide is intended by Brenner to show the downside of relying on economic aggregates and interventionism.
Bad Models, Bad Policy
It’s easy to see that economic aggregates have numerous flaws, rendering them unreliable guides for monetary and fiscal policy. Nevertheless, their publication has tended to encourage the adoption of policy interventions. This points to another issue lurking in the background: the role of economic aggregates in shaping the theory and practice of macroeconomics and the models on which policy recommendations are based. The conceptual difficulties surrounding aggregates, and the errors embedded within measured aggregates, have helped to foster questionable model treatments from a scientific perspective. For example, Paul Romer has said:
“Macroeconomists got comfortable with the idea that fluctuations in macroeconomic aggregates are caused by imaginary shocks, instead of actions that people take, after Kydland and Prescott (1982) launched the real business cycle (RBC) model. … [which] explains recessions as exogenous decreases in phlogiston.”
This is highly reminiscent of a quip by Brenner that macroeconomics has become a bit like astrology. A succession of macro models after the RBC model inherited the dependence on phlogiston. Romer goes on to note that model dependence on “imaginary” forces has aggravated the longstanding problem of statistically identifying individual effects. He also debunks the notion that adding expectations to models helps solve the identification problem. In fact, Romer insists that it makes it worse. He goes on to paint a depressing picture of the state of macroeconomics, one to which its reliance on faulty aggregates has surely contributed.
Aggregates also mask the detailed, real-world impacts of policies that invariably accompany changes in spending and taxes. While a given fiscal policy initiative might appear to be neutral in aggregate terms, it is almost always distortionary. For example, spending and tax programs always entail a redirection of resources, whether a consequence of redistribution, large-scale construction, procurement, or efforts to shape the industrial economy. These are usually accompanied by changes in the structure of incentives, regulatory requirements, and considerable rent seeking activity. Too often, outlays are dedicated to shoring up weak sectors of the economy, short-circuiting the process of creative destruction that serves to foster economic growth. Yet the macro models gloss over all the messy details that can negate the efficacy of activist fiscal policies.
Conclusion
The reliance of macroeconomic policy on aggregates like GDP, employment, and inflation statistics certainly has its dangers. These measures all suffer from theoretical problems, and they simply cannot be calculated without errors. They are backward-looking, and the necessity of making ongoing revisions leads to greater uncertainty. But compared to what? There are ways of shifting the focus to measures subject to less uncertainty, such as nominal income rather than real income. A number of theorists have proposed market-based methods of guiding policy, including Fischer Black. This deserves broader discussion.
The problems of aggregates are not solely confined to measurement. For example, national income accounting, along with the Keynesian focus on “underconsumption” during recessions, led to the fallacious view that spending decisions drive the economy. This became macroeconomic orthodoxy, driving macro mismanagement for decades and leading to inexorable growth in the dominance of government. Furthermore, macroeconomic models themselves have been corrupted by the effort to explain away impossibly error-prone measurements of aggregate activity.
Brenner has a point: it might be more productive to ignore the economic aggregates and institute stable policies which reinforce the efficacy of private markets in allocating resources. If nothing else, it makes sense to feature the government and private components separately.
Supporters of President Trump’s hard line on trade make so many false assertions that it’s hard to keep up. I’ve addressed several of these in earlier posts and I’ll address two more fallacies here: 1) that the U.S. manufacturing sector is in a state of crisis; and 2) that tariffs played a key role in promoting economic growth in the U.S. during the so-called gilded age of the late 19th and early 20th centuries.
Security
First, let’s revisit one tenet of protectionism: national security demands self-sufficiency. This undergirds the story that we must produce physical “things”, in addition to often higher-valued services, to be a great nation, or even to survive!
Of course, protecting industries critical to national security might seems like a natural concession to make, even for those supportive of liberalized trade. Ross Douthat says this:
“I think trying to reshore some manufacturing and decouple more from China makes sense from a national security standpoint, even if it costs something to G.D.P. and the stock market.“
Unfortunately, this kind of rationale is far too malleable. There is never a clearly defined limiting principle. Someone decides which goods are “critical” to national security, and this deliberation becomes the subject of much political jockeying and favor-seeking. But wait! Economic security is also cited as an adequate excuse for trade protections! And how about data security? Health security? Job security? Always there is insistence that “security” of one sort or another demands that we provide for our own needs. For definitive proof, take a look at this nonsense! Give them an inch and they’ll take a mile.
Pretty soon you “protect” such a wide swath of industries in a quest for self-sufficiency that the entire economy is unmoored from opportunity costs, comparative advantages, and the information about scarcities provided by market prices. Absolute “security” comes at the cost of transforming the economy’s productive machinery into a complacent hulk rivaling the inefficiency of Soviet industrial planning. Competition is the solution, but not limited to firms under the same set of protective trade barriers.
Manufacturing Is Mostly Fine
Trade warriors, including members of Trump’s team, insist that our decline as a nation is being hastened by a crisis in manufacturing. However, value added in U.S. manufacturing is at an all-time high.
There has been a long-term decline in manufacturing employment, but not manufacturing output. In fact, manufacturing output has doubled since 1980. As Jeff Jacoby notes, “the purpose of manufacturing is to make things, not jobs.” If our overarching social goal was job security, we’d have revolted long ago against the tremendous reduction in agricultural employment experienced over the past century. We’d rely on switchboard operators to load web pages, and we’d dig trenches and tunnels with spoons (to paraphrase Milton Friedman).
The secular decline in manufacturing employment is a consequence of growth in manufacturing productivity. Economy-wide, this phenomenon allows real income and our standard of living to grow.
Take That Job and …
It’s also significant that few Americans have much interest in factory work. It’s typically less dangerous than in times past, but many of today’s factory jobs are still physically challenging and relatively risky. Perhaps that helps explain why nearly half-a-million jobs in manufacturing are unfilled.
Jacoby describes the transition that has changed the face of American manufacturing:
“… US plants have largely turned away from making many of the low-tech, labor-intensive consumer items they once specialized in — sneakers, T-shirts, small appliances, toys. Those jobs have mostly gone overseas, and trying to bring them back by means of a trade war would be ruinous. Yet America remains a global manufacturing powerhouse — highly skilled, highly innovative, and highly efficient.“
And yet, even as wages in manufacturing have grown, many factory jobs do not pay as well as positions requiring far less strenuous toil in the services sector. It’s also true that the best manufacturing jobs in the U.S. today require high-level skills, which are in short supply. These factors help explain why manufacturers believe finding qualified workers is one of their biggest challenges.
Isolating Weak Sectors
There are specific sectors within manufacturing that have fared poorly, including textiles, furniture, metals, and low-end electronics. The loss of competitiveness that drove those sectoral declines is not a new development. It has, however, devastated communities in the U.S. that were heavily dependent on these industries. These misfortunes are regrettable, but trade barriers are not an effective prescription for revitalizing depressed areas.
Meanwhile, other manufacturing sectors have enjoyed growth, such as computers, aerospace, and EVs. While we’ve seen a decline in the number of manufacturing firms, theperformance of U.S. manufacturing in the 21st century can be described as mixed at the very worst.
The author of this piece seems to accept the false notion that U.S. manufacturing is moribund, but he knows tariffs aren’t an effective way to strengthen domestic goods production. He has a number of better suggestions, including a commitment to infrastructure investment, reforms to education and health, and reconfiguring certain corporate income tax policies. Unfortunately, his ideas on tariffs are sometimes as mistaken as Trump’s,
The Gilded Age
Finally, the other false assertion noted in the opening paragraph is that tariffs somehow spurred economic growth in the late 19th and early 20th centuries. Brian Albrecht corrects this protectionist fallacy, which lies at the root of many defenses of Trump’s tariffs. Albrecht cites favorable conditions for growth that were sufficient to overwhelm the negative effects of tariffs, including:
“… explosive population growth, mass European immigration, rapid technological innovation, westward expansion, abundant natural resources, high literacy rates, and stable property rights.”
While cross-country comparisons indicate a positive correlation between tariffs and growth during the 1870 – 1920 period, those differences were caused by other forces that dominated tariffs. Cross-industry research discussed by Albrecht indicates that tariffs on manufactured goods during the gilded era reduced labor productivity and stimulated the entry of smaller, less productive firms. Likewise, natural experiments find that tariffs allowed inefficient firms to survive and discouraged innovation.
Conclusion
The U.S. manufacturing sector is not in any sort of crisis, and its future growth won’t be powered by attempts to restore the sort of low-value production offshored over the past several decades. What protectionists interpret as failure is the natural progression of a technically advanced market-based civilization, where high-value services account for greater shares of growing total output. Of course, low-value production is sometimes “crowded out” in this process, depending on its trade-ability and comparative advantages. The logic of the process is encapsulated by Veronique de Rugy’s recent discussion of iPhone production (HT: Don Boudreaux):
“Then there’s [Commerce Secretary Howard] Lutnick, pining for a world where Americans flood back into massive factories to assemble iPhones. This is nostalgic industrial cosplay masquerading as economic strategy. Yes, iPhones aren’t assembled by Americans. But this isn’t a failure; it’s a feature of smart economic specialization. We design the iPhone here. That’s the high-value, high-margin part. The sophisticated chips, software, architecture, and intellectual property are all created in the U.S. The marketing is done here, too. That’s most of the value of the iPhone. The lower-value labor-intensive assembly work is done abroad because those tasks are more efficiently performed abroad.“
There is certainly no crisis in U.S. manufacturing. That narrative is driven by a combination of politics, rent seeking, and misplaced nostalgia.
Matt Yglesias tweeted on X that “the bond market does not appear to believe in DOGE”. He included a chart much like the updated one above to “prove” his point. Tyler Cowen posted a link to the tweet on Marginal Revolution, without comment … Cowen surely must know that any such conclusion is premature, especially based on the movement of Treasury yields over the past month (or more, since the market’s evaluation of the DOGE agenda preceded Trump’s inauguration).
Of course, there is a difference between “believing” in DOGE and being convinced that its efforts should have succeeded in reducing interest rates immediately amidst waves of background noise from budget and tax legislation, court challenges, Federal Reserve missteps (this time cutting rates too soon), and the direction of the economy in general.
In this case, perhaps a better way to define success for DOGE is a meaningfully negative impact on the future supply of Treasury debt. Even that would not guarantee a decline in Treasury rates, so the premise of Yglesias’ tweet is somewhat shaky to begin with. Still, all else equal, we’d expect to see some downward pressure on yields if DOGE succeeds in this sense. But we must go further by recognizing that DOGE savings could well be reallocated to other spending initiatives. Then, the savings would not translate into lower supplies of Treasury debt after all.
Certainly, the DOGE team has made progress in identifying wasteful expenditures, inefficiencies, and poor controls on spending. But even if the $55 billion of estimated savings to date is reliable, DOGE has a long way to go to reach Musk’s stated objective of $2 trillion. There are some juicy targets, but it will be tough to get there in 17 more months, when DOGE is to stand down. Still, it’s not unreasonable to think DOGE might succeed in accomplishing meaningful deficit reduction.
But if bond traders have doubts about DOGE, it’s partly because Donald Trump and Elon Musk themselves keep giving them reasons. In my view, Musk and Trump have made a major misstep in toying with the idea of using prospective DOGE savings to fund “dividend checks” of $5,000 for all Americans. These would be paid by taking 20% of the guesstimated $2 trillion of DOGE savings. Musk’s expression of interest in the idea was followed by a bit of clusterfuckery, as Musk walked back his proposal the next day even as Trump jumped on board. PLEASE Elon, don’t give the Donald any crowd-pleasing ideas! And don’t lose sight of the underlying objective to reduce the burden of government and the public debt.
Now, Trump proposes that 60% of the savings accomplished by DOGE be put toward paying for outlays in future years. Sure, that’s deficit reduction, but it may serve to dull the sense that shrinking the federal government is an imperative. The mechanics of this are unclear, but as a first pass, I’d say the gain from investing DOGE savings for a year in low-risk instruments is unlikely to outweigh the foregone savings in interest costs from paying off debt today! Of course, that also depends on the future direction of interest rates, but it’s not a good bet to make with public funds.
Nor can the bond market be comforted by uncertainty surrounding legislation that would not only extend the Trump tax cuts, but will probably include various spending provisions, both cuts and increases. As of now, the mix of provisions that might accompany a deal among GOP factions is very much up in the air.
There is also trepidation about Trump’s aggressive stance toward the Federal Reserve. He promises to replace Jerome Powell as Fed Chairman, but with God knows whom? And Trump jawbones aggressively for lower rates. The Fed’s ill-advised rate cuts in the fall might have been motivated in part by an attempt to capitulate to the then-President Elect.
Trump’s Executive Order to create a sovereign wealth fund (SWF), which I recently discussed here, is probably not the most welcome news to bond investors. All else equal, placing tax or tariff revenue into such a fund would reduce the potential for deficit reduction, to say nothing of the idiocy of additional borrowing to purchase assets.
Finally, Trump has proposed what might later prove to be massive foreign policy trial balloons. Some of these are bound up with the creation of the SWF. They might generate revenue for the government without borrowing (mineral rights in Ukraine? Or Greenland?), but at this point there’s also a chance they’ll create massive funding needs (Gaza development?). Again, Trump seems to be prodding or testing counterparties to various negotiations… prodding diplomacy. It’s unlikely that anything too drastic will come of it from a fiscal perspective, but it probably doesn’t leave bond traders feeling easy.
At this stage, it’s pretty rash to conclude that the bond market “doesn’t believe in DOGE”. In fact, there is no doubt that DOGE is making some progress in identifying potential fraud and inefficiencies. However, bond traders must weigh a wide range of considerations, and Donald Trump has a tendency to kick up dust. Indeed, the so-called DOGE dividend will undermine confidence in debt reduction and bond prices.
Kamala Harris’ campaign platform lifts several tax provisions from Joe Biden’s ill-fated campaign. The most pernicious of these are lauded by observers on the Left for their “fairness”, but they dismiss some rather obvious economic damage these provisions would inflict. Here, I’ll cover Harris’ proposal to tax unrealized capital gains of the rich in two different ways:
A minimum 25% “billionaire tax” on the “incomes” of taxpayers with net worth exceeding $100 million. This definition of income would include unrealized capital gains.
A tax of 28% at the time of death on unrealized capital gains in excess of $5 million ($10 million for joint returns).
Why Bother?
To get a whiff of the complexity involved, take a look at the description on pp. 79 – 85 of this document, to which the Harris proposal seems to correspond. It’s not fully fleshed out, but it’s easy to imagine the lucrative opportunities this would create for tax attorneys and accountants, to say nothing of job openings at the IRS!
On the other hand, there’s little chance these proposals would be approved by Congress, no matter which party holds a majority. Harris knows that, or at least her advisors do. That taxation of unrealized gains is even part of the conversation in a presidential election year tells us how normalized the idea has become within the Democrat Party, which seems to have lost all regard for private property rights. These are classist proposals designed to garner the votes of the “tax-the-rich” crowd, who either aren’t aware or haven’t come to grips with the fact that the U.S. already has a very progressive income tax system. “The rich” already pay a disproportionately high share of taxes.
Taxable Income
These provisions would complicate and corrupt the income tax code by distorting the definition of income for tax purposes. The Internal Revenue Code has always been consistent in defining taxable income as realized income. One might use the expression “mark-to-market taxation” to characterize a tax on unrealized gains from tradable assets. It’s much more difficult to estimate unrealized gains on non-tradable or infrequently traded investments, for which there is no ready market value.
There is one type of income that some believe to be taxed as unrealized. A few weeks ago, in a post about Sam Altman’s infatuation with a wealth tax, I cited a recent Supreme Court decision that has been mistakenly interpreted as favoring income taxation of unrealized gains or a wealth tax. In fact, Moore v. United States involved the undistributed profits of a foreign pass-through entity (i.e., not a C corporation) for purposes of the mandatory repatriation tax. The foreign firm’s profits were realized, and its pass-through status meant that the U.S. owners had also, by definition, realized the profits. So this case did not set a precedent or create an exception to the rule that income taxation applies only to realized income.
Forced Sales
Tradable assets with easily recorded market values will often have unrealized gains in a given year. While tax payments might be spread over the current and future tax years, these taxes could necessitate asset sales to pay the taxes owed. If unrealized losses are treated symmetrically, they would require either future deductions or possibly credits for prior tax payments.
Estimates of unrealized gains on illiquid or private investments like closely-held business interests, artwork, or real estate are highly uncertain and subject to dispute. A large tax liability on such an asset could be especially burdensome. Cash must be raised, which might require a forced sale of other assets. And again, these valuations often come with great complexity and exorbitant administrative costs, not just for the IRS, but especially for taxpayers.
Economic Downsides
As I noted above, additional taxes on unrealized gains would create an obvious need for liquidity, if not immediately then at death. With or without careful planning, sales of assets by wealthy investors to pay the tax would undermine market values of equity (and other assets), producing a broader loss of wealth economy-wide.
Avoidance schemes would be heavily utilized. For example, a wealthy investor could borrow heavily against assets so as to offset unrealized gains with deductible debt-service costs.
Capital flight is likely to be intense if a Harris tax regime began to take shape in Congress. This might be the best avoidance scheme of all. The U.S. is likely to experience massive capital outflows. Furthermore, investment in new physical capital will decline, ultimately leading to lower productivity and real wages.
Entrepreneurial activity would also take a hit. In a critique of Jason Furman’s effort to justify Harris’ proposal, Tyler Cowen asks why we should be so eager to “whack” venture capital. He also quotes an email from Alex Tabarrok on the detrimental policy effects on rapidly growing start-ups:
“What’s really going on is that you are divorcing the entrepreneur from their capital at precisely the moment that the team is likely most productive. Separation of capital from entrepreneur could negatively impact the company’s growth or the entrepreneur’s ability to manage effectively. The entrepreneur could lose control, for example. If you wait until the entrepreneur realizes the gain that’s the time that the entrepreneur wants out and is ready to consume so it’s closer to taxing consumption and better timed in the entrepreneurial growth process.“
Or the entrepreneur might just decide that a startup would be more rewarding in a tax-friendly environment, perhaps somewhere overseas.
Interest Rates and Tax Receipts
Tabarrok notes in a separate post that much of the variation in stock prices is caused by changes in interest rates. Investors use market rates to determine discount rates at which a firm’s future cash flows can be valued. Thus, changes in rates engender changes in stock prices, capital gains, and capital losses.
A decline in interest rates can raise market valuations without any change in dividends. However, a long-term investor would see no change in pre-tax income or consumption, so the tax could force a series of premature sales. A change in a firm’s expected growth rate would also create an unrealized gain (or loss), but the tax would undermine U.S. equity values. Taxing an actual increase in the dividend is one thing, but taxing a change in expectations of future dividends is another. As Tabarrok puts it, “It’s taxing the chickens before the eggs have hatched.“
Dangerous Narrative, Dangerous Policy
A final objection to taxing unrealized capital gains is that it would cross the line into a form of wealth taxation. Assets come in many forms, but the only time realized values can be discerned are when they are traded. That goes for collectibles, homes, boats, and the full array of financial assets. A corollary is that a very large percentage of wealth is unrealized.
A tax on unrealized gains would be the proverbial camel’s nose under the tent and another incursion into the private realm. So often in the history of taxation we’ve seen narrow taxes expand into broad taxes. This is one more opportunity for the state to extend its dominance and control.
I’ve written in the past about the economic dangers of a wealth tax. First, every dollar of income used to purchase capital is already taxed once. In that sense, the cost basis of wealth would be double taxed under a wealth tax. Second, the supply of capital is highly elastic. This implies a high propensity for capital flight, shallowing of productive physical capital, and reduced productivity and real wages. Avoidance schemes would rapidly be put into play. Given these limitations, the revenue raising potential of a wealth tax is unlikely to live up to expectations. Finally, a wealth tax is unconstitutional, but that won’t stop the Left from pushing for one, especially if they first get a tax on unrealized gains. Even if they are unsuccessful now, the conversation tends to normalize the idea of a wealth tax among low-information voters, and that is a shame.
I was happy to see Noah Smith’s recent post on the graces of comparative advantage and the way it should mediate the long-run impact of AI on job prospects for humans. However, I’m embarrassed to have missed his post when it was published in March (and I also missed a New York Timespiece about Smith’s position).
I said much the same thing as Smith in my post two weeks ago about the persistence of a human comparative advantage, but I wondered why the argument hadn’t been made prominently by economists.I discussed it myself about seven years ago. But alas, I didn’t see Smith’s post until last week!
I highly recommend it, though I quibble on one or two issues. Primarily, I think Smith qualifies his position based on a faulty historical comparison. Later, he doubles back to offer a kind of guarantee after all. Relatedly, I think Smith mischaracterizes the impact of energy costs on comparative advantages, and more generally the impact of the resources necessary to support a human population.
We Specialize Because…
Smith encapsulates the underlying phenomenon that will provide jobs for humans in a world of high automation and generative AI: “… everyone — every single person, every single AI, everyone — always has a comparative advantage at something!” He tells technologists “… it’s very possible that regular humans will have plentiful, high-paying jobs in the age of AI dominance — often doing much the same kind of work that they’re doing right now …”
… often, but probably transformed in fundamental ways by AI, and also doing many other new kinds of work that can’t be foreseen at present. Tyler Cowen believes themost important macro effects of AI will be from “new” outputs, not improvements in existing outputs. That emphasis doesn’t necessarily conflict with Smith’s narrative, but again, Smith thinks people will do many of the same jobs as today in a world with advanced AI.
Smith’s Non-Guarantee
Smith hedges, however, in a section of his post entitled “‘Possible’ doesn’t mean guaranteed”. This despite his later assertion that superabundance would not eliminate jobs for humans. That might seem like a separate issue, but it’s strongly intertwined with the declining AI cost argument at the basis of his hedge. More on that below.
On his reluctance to “guarantee” that humans will have jobs in an AI world, Smith links to a 2013 Tyler Cowen post on“Why the theory of comparative advantage is overrated”. For example, Cowen says, why do we ever observe long-term unemployment if comparative advantage rules the day? Of course there are many reasons why we observe departures from the predicted results of comparative advantage. Incentives are often manipulated by governments and people differ drastically in their capacities and motivation.
But Cowen cites a theoretical weakness of comparative advantage: that inputs are substitutable (or complementary) by degrees, and the degree might change under different market conditions. An implication is that “comparative advantages are endogenous to trade”, specialization, and prices. Fair enough, but one could say the same thing about any supply curve. And if equilibria exist in input markets it means these endogenous forces tend toward comparative advantages and specializations balancing the costs and benefits of production and trade. These processes might be constrained by various frictions and interventions, and their dynamics might be complex and lengthy, but that doesn’t invalidate their role in establishing specializations and trade.
The Glue Factory
Smith concerns himself mainly with another one of Cowen’s “failings of comparative advantage”: “They do indeed send horses to the glue factory, so to speak.” The gist here is that when a new technology, motorized transportation, displaced draft horses, there was no “wage” low enough to save the jobs performed by horses. Smith says horses were too costly to support (feed, stables, etc…), so their comparative advantage at “pulling things” was essentially worthless.
True, but comparing outmoded draft horses to humans in a world of AI is not quite appropriate. First, feedstock to a “glue factory” better not be an alternative use for humans whose comparative advantages become worthless. We’ll have to leave that question as an imperative for the alignment community.
Second, horses do not have versatile skill sets, so the comparison here is inapt due to their lack of alternative uses as capital assets. Yes, horses can offer other services (racing, riding, nostalgic carriage rides), but sadly, the vast bulk of work horses were “one-trick ponies”. Most draft horses probably had an opportunity cost of less than zero, given the aforementioned costs of supporting them. And it should be obvious that a single-use input has a comparative advantage only in its single use, and only when that use happens to be the state-of-the-art, or at least opportunity-cost competitive.
The drivers, on the other hand, had alternatives, and saw their comparative advantage in horse-driving occupations plunge with the advent of motorized transport. With time it’s certain many of them found new jobs, perhaps some went on to drive motorized vehicles. The point is that humans have alternatives, the number depending only on their ability to learn a crafts and perhaps move to a new location. Thus, as Smith says, “… everyone — every single person, every single AI, everyone — always has a comparative advantage at something!” But not draft horses in a motorized world, and not square pegs in a world of round holes.
AI Producer Constraints
That brings us to the topic of what Smith calls producer-specific constraints, which place limits on the amount and scope of an input’s productivity. For example, in my last post, there was only one super-talented Harvey Specter, so he’s unlikely to replace you and keep doing his own job. Thus, time is a major constraint. For Harvey or anyone else, the time constraint affects the slope of the tradeoff (and opportunity costs) between one type of specialization versus another.
Draft horses operated under the constraints of land, stable, and feed requirements, which can all be viewed as long-run variable costs. The alternative use for horses at the glue factory did not have those costs.
Humans reliant on wages must feed and house themselves, so those costs also represent constraints, but they probably don’t change the shape of the tradeoff between one occupation and another. That is, they probably do not alter human comparative advantages. Granted, some occupations come with strong expectations among associates or clients regarding an individual’s lifestyle, but this usually represents much more than basic life support. In the other end of the spectrum, displaced workers will take actions along various margins: minimize living costs; rely on savings; avail themselves of charity or any social safety net as might exist; and ultimately they must find new positions at which they maintain comparative advantages.
The Compute Constraint
In the case of AI agents, the key constraint cited by Smith is “compute”, or computer resources like CPUs or GPUs. Advancements in compute have driven the AI revolution, allowing AI models to train on increasingly large data sets and levels of compute. In fact, by one measure of compute, floating point operations per second (FLOPs), compute has become drastically cheaper, with FLOPs per dollar almost doubling every two years. Perhaps I misunderstand him, but Smith seems to assert the opposite: that compute costs are increasing. Regardless, compute is scarce, and will always be scarce because advancements in AI will require vast increases in training. This author explains that while lower compute costs will be more than offset by exponential increases in training requirements, there nevertheless will be an increasing trend in capabilities per compute.
Every AI agent will require compute, and while advancements are enabling explosive growth in AI capabilities, scarce compute places constraints on the kinds of AI development and deployment that some see as a threat to human jobs. In other words, compute scarcity can change the shape of the tradeoffs between various AI applications and thus, comparative advantages.
The Energy Constraint
Another producer constraint on AI is energy. Certainly highly complex applications, perhaps requiring greater training, physical dexterity, manipulation of materials, and judgement, will require a greater compute and energy tradeoff against simpler applications. Smith, however, at one point dismisses energy as a differential producer constraint because “… humans also take energy to run.” That is a reference to absolute energy requirements across inputs (AI vs. human), not differential requirements for an input across different outputs. Only the latter impinge on tradeoffs or opportunity costs facing an inputs. Then, the input having the lowest opportunity cost for a particular output has a comparative advantage for that output. However, it’s not always clear whether an energy tradeoff across outputs for humans will be more or less skewed than for AI, so this might or might not influence a human comparative advantage.
Later, however, Smith speculates that AI might bid up the cost of energy so high that “humans would indeed be immiserated en masse.” That position seems inconsistent. In fact, if AI energy demands are so intensive, it’s more likely to dampen the growth in demand for AI agents as well as increase the human comparative advantage because the most energy-intensive AI applications will be disadvantaged.
And again, there is Smith’s caution regarding the energy required for human life support. Is that a valid long-run variable cost associated with comparative advantages possessed by humans? It’s not wrong to include fertility decisions in the long-run aggregate human labor supply function in some fashion, but it doesn’t imply that energy requirements will eliminate comparative advantages. Those will still exist.
Hype, Or Hyper-Growth?
AI has come a long way over the past two years, and while its prospective impact strikes some as hyped thus far, it has the potential to bring vast gains across a number of fields within just a few years. According to this study, explosive economic growth on the order of 30% annually is a real possibility within decades, as generative AI is embedded throughout the economy. “Unprecedented” is an understatement for that kind of expansive growth. Dylan Matthews in Vox surveys the arguments as to how AI will lead to super-exponential economic growth. This is the kind of scenario that would give rise to superabundance.
I noted above that Smith, despite his unwillingness to guarantee that human jobs will exist in a world of generative AI, asserts (in an update) at the bottom of his post that a superabundance of AI (and abundance generally) would not threaten human comparative advantages. This superabundance is a case of decreasing costs of compute and AI deployment. Here Smith says:
“The reason is that the more abundant AI gets, the more value society produces. The more value society produces, the more demand for AI goes up. The more demand goes up, the greater the opportunity cost of using AI for anything other than its most productive use.
“As long as you have to make a choice of where to allocate the AI, it doesn’t matter how much AI there is. A world where AI can do anything, and where there’s massively huge amounts of AI in the world, is a world that’s rich and prosperous to a degree that we can barely imagine. And all that fabulous prosperity has to get spent on something. That spending will drive up the price of AI’s most productive uses. That increased price, in turn, makes it uneconomical to use AI for its least productive uses, even if it’s far better than humans at its least productive uses.
“Simply put, AI’s opportunity cost does not go to zero when AI’s resource costs get astronomically cheap. AI’s opportunity cost continues to scale up and up and up, without limit, as AI produces more and more value.”
This seems as if Smith is backing off his earlier hedge. Some of that spending will be in the form of fabulous investment projects of the kinds I mentioned in my post, and smaller ones as well, all enabled by AI. But the key point is that comparative advantages will not go away, and that means human inputs will continue to be economically useful.
I referenced Andrew Mayne in my last post. He contends that the income growth made possible by AI will ensure that plenty of jobs are available for humans. He mentions comparative advantage in passing, but he centers his argument around applications in which human workers and AI will be strong complements in production, as will sometimes be the case.
A New Age of Worry
The economic success of AI is subject to a number of contingencies. Most important is that AI alignment issues are adequately addressed. That is, the “self-interest” of any agentic AI must align with the interests of human welfare. Do no harm!
The difficulty of universal alignment is illustrated by the inevitability of competition among national governments for AI supremacy, especially in the area of AI-enabled weaponry and espionage. The national security implications are staggering.
A couple of Smith‘s biggest concerns are the social costs of adjusting to the economic disruptions AI is sure to bring, as well as its implications for inequality. Humans will still have comparative advantages, but there will be massive changes in the labor market and transitions that are likely to involve spells of unemployment and interruptions to incomes for some. The speed and strength of the AI revolution may well create social upheaval. That will create incentives for politicians to restrain the development and adoption of AI, and indeed, we already see the stirrings of that today.
Finally, Smith worries that the transition to AI will bring massive gains in wealth to the owners of AI assets, while workers with few skills are likely to languish. I’m not sure that’s consistent with his optimism regarding income growth under AI, and inequality matters much less when incomes are rising generally. Still, the concern is worthy of a more detailed discussion, which I’ll defer to a later post.
I’ve taken an extended hiatus from blogging while moving to a different part of the country. I haven’t posted here in over 10 weeks, but a new post appears below. I’m still tying-up loose ends from the move, but I’ll be trying to get back to posting more regularly … trying!
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Absurd ideas about race and identity politics come from extreme elements on both the Left and the Right. Some leftists insist that race has no natural basis — that it’s simply a “social construct”. On the Right, a “racialist” contingent is promoting the “celebration of whiteness” and embracing racial preferences for whites. Treated as alternative pathways, I’d take “social construct”. It’s nonsense, of course, but the beautiful irony is that it provides a basis for stripping away from our institutions the entire diversity, equity, and inclusion (DEI) straightjacket. It’s almost as if those promoting race as a social construct wish to build a “colorblind” society. On the other hand, I suppose some think they can have their DEI cake along with a side of free choice to identify as anything they want: black, white, or furry.
Who Are the Racists?
People of good faith don’t harbor or act on racist tendencies. The mere recognition of racial/ethnic/cultural differences is not evidence of racism and does not preclude the treatment of all with fairness and due respect. It’s possible to respect, value, or fall in love with someone outside one’s own racial, ethnic, or cultural group of origin, even while holding a general affinity for one’s own group, as nearly everyone does.
But a few real racists are sprinkled across all races, ethnicities, cultures, and the full political spectrum. The “popular” racist stereotype as white male has been kept alive by the lingering echos of slavery in America, which ended nearly 16 decades ago, and the long hangover that included Jim Crow laws and segregation. Today, however, “white society” or “whiteness” is hardly the sole domain of prejudice.
IsRacialism Different?
Now, a few whites are promoting the celebration of “white identity” as a counterbalance to identity politics among non-whites. Ostensibly, this “white racialism” might be similar to celebrations of identity often practiced by minorities, which are also forms of racialism. Should white racialism be viewed as less savory than racialism practiced by racial minorities?
For most Caucasians, “being white” does not have much salience relative to other affiliations defining identity. That’s why white racialism seems odd to me. Sure, when forced to check a box, whites will check “Caucasian”, but “white identity” seems overly broad. There are too many distinct cultures and subcultures that dominate self-identity, such as national ancestry, religion, and cultural membership.
The same could be said for many other racial categories, but minority status and historical events (e.g., American slavery) help explain why broad categories often form cohesive identity groups. And, as Christopher Rufo notes in his great discussion of the racialist viewpoint, broad categories tend to be the most closely associated with racialism:
“Yes, left-wing racialism is indeed now deeply embedded in America’s institutions, and the demographic balance of the country has shifted in recent decades. And yes, the basic racial classification system in the United States broadly delineates continental origin—Europe, Africa, Latin America, Asia—in a way that is not arbitrary or meaningless. Terms such as ‘white,’ ‘black,’ ‘Latino,’ and ‘Asian,’ while often obscuring important variations within such groupings, have become the lingua franca and are useful shorthand descriptors for many purposes.”
There are individuals from all groups or “classes”, including whites, who react critically to aggressive expressions of identity by members of other classes. Perhaps that’s excusable, depending on the degree of zealotry on either part. The line between pride in race/ancestry/culture and fractious racialism might be hard to discern in some cases, but the chief distinction is rooted in explicit, demeaning and/or envious comparisons to “out-groups”. This might be damaging enough, but from there it can be a very short step into outright racism.
A preoccupation with the historic disadvantages of one’s race can be disempowering to an individual and destructive in a social sense. I believe the white racialist phenomenon belongs in that category. The presumed “disadvantages” of whiteness are very contemporary, however, rooted in policies dating back only to the widespread adoption of racial preferences for non-white “protected classes” and DEI.
Preferences For All
Imagine the racialist policies now practiced widely in government, industry, and academia — particularly racial preferences on behalf of protected classes — but now applied on behalf of heretofore unprotected classes as well. For example, what if some proportion of jobs, admissions, or other coveted placements were set aside for whites? If whites represent 50% of the population, then 50% of hires or admissions would be reserved for whites.
Some might assume that this treatment is already implied by existing racial preferences, but that’s not the case. In the wake of George Floyd’s death, just 6% of new hires among S&P companies were white, according to Bloomberg News.
Nevertheless, such a white racialist turnabout would be a colossal mistake. Adding strict limits to the application of existing preferences might be a good thing, but white racial preferences would buttress the entire system of racial preferences as an institution and add more rigidity to the operation of labor markets. From an economic viewpoint, it would be just as pernicious as racial preferences generally.
Racial preferences of any kind freeze labor markets and impair the allocation of human resources to their most-valued uses. In fact, placing one individual into a position on any basis other than their qualifications implies that two individuals must be placed into positions in which they lack comparative advantage relative to each other. Little by little, that means lost output and upward price pressure. It is a mechanism that short circuits gains from trade, shriveling the benefits that the most and least talented confer on society at large. Extending preferences to whites would only serve to further institutionalize this damaging practice.
Adherence to numerical preferences is to pretend that people can be treated less as individuals and more like interchangeable parts… except with respect to their value as “class members”. Racial preferences are presumed to be a remedy for so-called structural racism, as opposed to racism by individuals. But they involve classification and favor the so-called “oppressed” at the expense of designated “oppressors”. The latter, almost without exception, had no role oppressive regimes of the past. Favoritism of this kind necessarily means reverse discrimination and fails to match individuals to roles in an optimal fashion.
Whether publicly or privately imposed, racial preferences often undermine those they are purported to help by placing individuals into positions for which they may not be competitive. This can sabotage an individual’s long-term success. It goes without saying that preferences build resentment among the “unprotected”, which goes to the impetus for “white racialism”. Indeed, preferences are not always popular with protected classes either. That’s because they interfere with merit-based decision-making and are perceived to stigmatize those presumed to benefit.
The Fixed Pie Is a Lie
Racialism reflects zero-sum thinking, a hallmark of DEI initiatives. Tyler Cowen quoted the abstract of a recent NBER working paper that found:
“… a more zero-sum mindset is strongly associated with more support for government redistribution, race- and gender-based affirmative action, and more restrictive immigration policies.”
Zero-sum thinking is fundamental to rent-seeking behavior, which is motivated by either malevolent greed or perceptions of victimhood. Victimhood and rent seeking is at the heart of calls for DEI, to say nothing of more radical proposals like reparation payments. White racialism attempts to get in on the action by positing that whites are oppressed under the current institutional dominance of DEI. But the misguided presumption that every identity group should have their own preferences or quotas broadens the emphasis on redressing perceived harms and redistributing rewards — zero-sum activities.
These zero-sum efforts waste energy and resources, harming our ability to produce things that enhance well being. Ultimately, they are actually negative-sum activities, and they also breed hatred.
Race is obviously determined by genetics, but I’d be happy to pretend it’s a mere social construct if that would help get us to a “colorblind” society.
Conclusion
There’s a huge irony in the racialism exercised by both traditional and “white racialist” DEI advocates: it neglects the most fundamental and just application of diversity: equality of opportunity. This principle incorporates the concept of diversity without sacrificing economic efficiency. We’ve largely abandoned it in favor of equality of outcomes via racial preferences, even at a time when society has become enlightened with respect to racial differences. In doing so, we’ve unintentionally chosen another form of explicit racial victimization.
To close, here’s a good summary of the dangers of racialism and identity politics offered by Victor Davis Hanson:
“Anytime one ethnic, racial, or religious group refuses to surrender its prime identity in exchange for a shared sense of self, other tribes for their own survival will do the same.
All then rebrand their superficial appearance as essential not incidental to whom they are.
And like nuclear proliferation that sees other nations go nuclear once a neighboring power gains the bomb, so too the tribalism of one group inevitably leads only to more tribalism of others. The result is endless Hobbesian strife.”
And that’s how white racialism fits right in with the pernicious politics of identity. When you can, vote for the elimination, or at least reform, of DEI policies and practices, not for a reinforcement of identity politics.
Policy activists have long maintained that manipulating government policy can stabilize the economy. In other words, big spending initiatives, tax cuts, and money growth can lift the economy out of recessions, or budget cuts and monetary contraction can prevent overheating and inflation. However, this activist mirage burned away under the light of experience. It’s not that fiscal and monetary policy are powerless. It’s a matter of practical limitations that often cause these tools to be either impotent or destabilizing to the economy, rather than smoothing fluctuations in the business cycle.
The macroeconomics classes seem like yesterday: Keynesian professors lauded the promise of wise government stabilization efforts: policymakers could, at least in principle, counter economic shocks, particularly on the demand side. That optimistic narrative didn’t end after my grad school days. I endured many client meetings sponsored by macro forecasters touting the fine-tuning of fiscal and monetary policy actions. Some of those economists were working with (and collecting revenue from) government policymakers, who are always eager to validate their pretensions as planners (and saviors). However, seldom if ever do forecasters conduct ex post reviews of their model-spun policy scenarios. In fairness, that might be hard to do because all sorts of things change from initial conditions, but it definitely would not be in their interests to emphasize the record.
In this post I attempt to explain why you should be skeptical of government stabilization efforts. It’s sort of a lengthy post, so I’ve listed section headings below in case readers wish to scroll to points of most interest. Pick and choose, if necessary, though some context might get lost in the process.
Expectations Change the World
Fiscal Extravagance
Multipliers In the Real World
Delays
Crowding Out
Other Peoples’ Money
Tax Policy
Monetary Policy
Boom and Bust
Inflation Targeting
Via Rate Targeting
Policy Coordination
Who Calls the Tune?
Stable Policy, Stable Economy
Expectations Change the World
There were always some realists in the economics community. In May we saw the passing of one such individual: Robert Lucas was a giant intellect within the economics community, and one from whom I had the pleasure of taking a class as a graduate student. He was awarded the Nobel Prize in Economic Science in 1995 for his applications of rational expectations theory and completely transforming macro research. As Tyler Cowen notes, Keynesians were often hostile to Lucas’ ideas. I remember a smug classmate, in class, telling the esteemed Lucas that an important assumption was “fatuous”. Lucas fired back, “You bastard!”, but proceeded to explain the underlying logic. Cowen uses the word “charming” to describe the way Lucas disarmed his critics, but he could react strongly to rude ignorance.
Lucas gained professional fame in the 1970s for identifying a significant vulnerability of activist macro policy. David Henderson explains the famous “Lucas Critique” in the Wall Street Journal:
“… because these models were from periods when people had one set of expectations, the models would be useless for later periods when expectations had changed. While this might sound disheartening for policy makers, there was a silver lining. It meant, as Lucas’s colleague Thomas Sargent pointed out, that if a government could credibly commit to cutting inflation, it could do so without a large increase in unemployment. Why? Because people would quickly adjust their expectations to match the promised lower inflation rate. To be sure, the key is government credibility, often in short supply.”
Non-credibility is a major pitfall of activist macro stabilization policies that renders them unreliable and frequently counterproductive. And there are a number of elements that go toward establishing non-credibility. We’ll distinguish here between fiscal and monetary policy, focusing on the fiscal side in the next several sections.
Fiscal Extravagance
We’ve seen federal spending and budget deficits balloon in recent years. Chronic and growing budget deficits make it difficult to deliver meaningful stimulus, both practically and politically.
The next chart is from the most recent Congressional Budget Office (CBO) report. It shows the growing contribution of interest payments to deficit spending. Ever-larger deficits mean ever-larger amounts of debt on which interest is owed, putting an ever-greater squeeze on government finances going forward. This is particularly onerous when interest rates rise, as they have over the past few years. Both new debt is issued and existing debt is rolled over at higher cost.
Relief payments made a large contribution to the deficits during the pandemic, but more recent legislation (like the deceitfully-named Inflation Reduction Act) piled-on billions of new subsidies for private investments of questionable value, not to mention outright handouts. These expenditures had nothing to do with economic stabilization and no prayer of reducing inflation. Pissing away money and resources only hastens the debt and interest-cost squeeze that is ultimately unsustainable without massive inflation.
Hardly anyone with future political ambitions wants to address the growing entitlements deficit … but it will catch up with them. Social Security and Medicare are projected to exhaust their respective trust funds in the early- to mid-2030s, which will lead to mandatory benefit cuts in the absence of reform.
If it still isn’t obvious, the real problem driving the budget imbalance is spending, not revenue, as the next CBO chart demonstrates. The “emergency” pandemic measures helped precipitate our current stabilization dilemma. David Beckworth tweets that the relief measures “spurred a rapid recovery”, though I’d hasten to add that a wave of private and public rejection of extreme precautions in some regions helped as well. And after all, the pandemic downturn was exaggerated by misdirected policies including closures and lockdowns that constrained both the demand and supply sides. Beckworth acknowledges the relief measures “propelled inflation”, but the pandemic also seemed to leave us on a permanently higher spending path. Again, see the first chart below.
The second chart below shows that non-discretionary spending (largely entitlements) and interest outlays are how we got on that path. The only avenue for countercyclical spending is discretionary expenditures, which constitute an ever-smaller share of the overall budget.
We’ve had chronic deficits for years, but we’ve shifted to a much larger and continuing imbalance. With more deficits come higher interest costs, especially when interest rates follow a typical upward cyclical pattern. This creates a potentially explosive situation that is best avoided via fiscal restraint.
Putting other doubts about fiscal efficacy aside, it’s all but impossible to stimulate real economic activity when you’ve already tapped yourself out and overshot in the midst of a post-pandemic economic expansion.
Multipliers In the Real World
So-called spending multipliers are deeply beloved by Keynesians and pork-barrel spenders. These multipliers tell us that every dollar of extra spending ultimately raises income by some multiple of that dollar. This assumes that a portion of every dollar spent by government is re-spent by the recipient, and a portion of that is re-spent again by another recipient. But spending multipliers are never what they’re cracked up to be for a variety of reasons. (I covered these in“Multipliers Are For Politicians”, and also see this post.) There are leakages out of the re-spending process (income taxes, saving, imports), which trim the ultimate impact of new spending on income. When supply constraints bind on economic activity, fiscal stimulus will be of limited power in real terms.
If stimulus is truly expected to be counter-cyclical and transitory, as is generally claimed, then much of each dollar of extra government spending will be saved rather than spent. This is the lesson of the permanent income hypothesis. It means greater leakages from the re-spending stream and a lower multiplier. We saw this with the bulge in personal savings in the aftermath of pandemic relief payments.
Another side of this coin, however, is that cutting checks might be the government’s single-most efficient activity in execution, but it can create massive incentive problems. Some recipients are happy to forego labor market participation as long as the government keeps sending them checks, but at least they spend some of the income.
Delays
Another unappreciated and destabilizing downside of fiscal stimulus is that it often comes too late, just when the economy doesn’t need stimulus. That’s because a variety of delays are inherent in many spending initiatives: legislative, regulatory, legal challenges, planning and design, distribution to various spending authorities, and final disbursement. As I noted here:
“Even government infrastructure projects, heralded as great enhancers of American productivity, are often subject to lengthy delays and cost overruns due to regulatory and environmental rules. Is there any such thing as a federal ‘shovel-ready’ infrastructure project?”
Crowding Out
The supply of savings is limited, but when government borrows to fund deficits, it directly competes with private industry for those savings. Thus, funds that might otherwise pay for new plant, equipment, and even R&D are diverted to uses that should qualify as government consumption rather than long-term investment. Government competition for funds “crowds-out” private activity and impedes growth in the economy’s productive capacity. Thus, the effort to stimulate economic activity is self-defeating in some respects.
Other Peoples’ Money
Government doesn’t respond to price signals the way self-interested private actors do. This indifference leads to mis-allocated resources and waste. It extends to the creation of opportunities for graft and corruption, typically involving diversion of resources into uses that are of questionable productivity (corn ethanol, solar and wind subsidies).
Consider one other type of policy action perceived as counter-cyclical: federal bailouts of failing financial institutions or other troubled businesses. These rescues prop up unproductive enterprises rather than allowing waste to be flushed from the system, which should be viewed as a beneficial aspect of recession. The upshot is that too many efforts at economic stabilization are misdirected, wasteful, ill-timed, and pro-cyclical in impact.
Tax Policy
Like stabilization efforts on the spending side, tax changes may be badly timed. Tax legislation is often complex and can take time for consumers and businesses to adjust. In terms of traditional multiplier analysis, the initial impact of a tax change on spending is smaller than for expenditures, so tax multipliers are smaller. And to the extent that a tax change is perceived as temporary, it is made less effective. Thus, while changes in tax policy can have powerful real effects, they suffer from some of the same practical shortcomings for stabilization as changes in spending.
However, stimulative tax cuts, if well crafted, can boost disposable incomes and improve investment and work incentives. As temporary measures, that might mean an acceleration of certain kinds of activity. Tax increases reduce disposable incomes and may blunt incentives, or prompt delays in planned activities. Thus, tax policy may bear on the demand side as well as the timing of shifts in the economy’s productive potential or supply side.
Monetary Policy
Monetary policy is subject to problems of its own. Again, I refer to practical issues that are seemingly impossible for policy activists to overcome. Monetary policy is conducted by the nation’s central bank, the Federal Reserve (aka, the Fed). It is theoretically independent of the federal government, but the Fed operates under a dual mandate established by Congress to maintain price stability and full employment. Therein lies a basic problem: trying to achieve two goals that are often in conflict with a single policy tool.
Make no mistake: variations in money supply growth can have powerful effects. Nevertheless, they are difficult to calibrate due to “long and variable lags” as well as changes in money “velocity” (or turnover) often prompted by interest rate movements. Excessively loose money can lead to economic excesses and an overshooting of capacity constraints, malinvestment, and inflation. Swinging to a tight policy stance in order to correct excesses often leads to “hard landings”, or recession.
Boom and Bust
The Fed fumbled its way into engineering the Great Depression via excessively tight monetary policy. “Stop and go” policies in the 1970s led to recurring economic instability. Loose policy contributed to the housing bubble in the 2000s, and subsequent maladjustments led to a mortgage crisis (also see here). Don’t look now, but the inflationary consequences of the Fed’s profligacy during the pandemic prompted it to raise short-term interest rates in the spring of 2022. It then acted with unprecedented speed in raising rates over the past year. While raising rates is not always synonymous with tightening monetary conditions, money growth has slowed sharply. These changes might well lead to recession. Thus, the Fed seems given to a pathology of policy shifts that lead to unintentional booms and busts.
Inflation Targeting
The Fed claims to follow a so-called flexible inflation targeting policy. In reality, it has reacted asymmetrically to departures from its inflation targets. It took way too long for the Fed to react to the post-pandemic surge in inflation, dithering for months over whether the surge was “transitory”. It wasn’t, but the Fed was reluctant to raise its target rates in response to supply disruptions. At the same time, the Fed’s own policy actions contributed massively to demand-side price pressures. Also neglected is the reality that higher inflation expectations propel inflation on the demand side, even when it originates on the supply side.
Via Rate Targeting
At a more nuts and bolts level, today the Fed’s operating approach is to control money growth by setting target levels for several key short-term interest rates (eschewing a more direct approach to the problem). This relies on price controls (short-term interest rates being the price of liquidity) rather than allowing market participants to determine the rates at which available liquidity is allocated. Thus, in the short run, the Fed puts itself into the position of supplying whatever liquidity is demanded at the rates it targets. The Fed makes periodic adjustments to these rate targets in an effort to loosen or tighten money, but it can be misdirected in a world of high debt ratios in which rates themselves drive the growth of government borrowing. For example, if higher rates are intended to reduce money growth and inflation, but also force greater debt issuance by the Treasury, the approach might backfire.
Policy Coordination
While nominally independent, the Fed knows that a particular monetary policy stance is more likely to achieve its objectives if fiscal policy is not working at cross purposes. For example, tight monetary policy is more likely to succeed in slowing inflation if the federal government avoids adding to budget deficits. Bond investors know that explosive increases in federal debt are unlikely to be repaid out of future surpluses, so some other mechanism must come into play to achieve real long-term balance in the valuation of debt with debt payments. Only inflation can bring the real value of outstanding Treasury debt into line. Continuing to pile on new debt simply makes the Fed’s mandate for price stability harder to achieve.
Who Calls the Tune?
The Fed has often succumbed to pressure to monetize federal deficits in order to keep interest rates from rising. This obviously undermines perceptions of Fed independence. A willingness to purchase large amounts of Treasury bills and bonds from the public while fiscal deficits run rampant gives every appearance that the Fed simply serves as the Treasury’s printing press, monetizing government deficits. A central bank that is a slave to the spending proclivities of politicians cannot make credible inflation commitments, and cannot effectively conduct counter-cyclical policy.
Stable Policy, Stable Economy
Activist policies for economic stabilization are often perversely destabilizing for a variety of reasons. Good timing requires good forecasts, but economic forecasting is notoriously difficult. The magnitude and timing of fiscal initiatives are usually wrong, and this is compounded by wasteful planning, allocative dysfunction, and a general absence of restraint among political leaders as well as the federal bureaucracy..
Predicting the effects of monetary policy is equally difficult and, more often than not, leads to episodes of over- and under-adjustment. In addition, the wrong targets, the wrong operating approach, and occasional displays of subservience to fiscal pressure undermine successful stabilization. All of these issues lead to doubts about the credibility of policy commitments. Stated intentions are looked upon with doubt, increasing uncertainty and setting in motion behaviors that lead to undesirable economic consequences.
The best policies are those that can be relied upon by private actors, both as a matter of fulfilling expectations and avoiding destabilization. Federal budget policy should promote stability, but that’s not achievable institutions unable to constrain growth in spending and deficits. Budget balance would promote stability and should be the norm over business cycles, or perhaps over periods as long as typical 10-year budget horizons. Stimulus and restraint on the fiscal side should be limited to the effects of so-called automatic stabilizers, such as tax rates and unemployment compensation. On the monetary side, the Fed would do more to stabilize the economy by adopting formal rules, whether a constant rate of money growth or symmetric targeting of nominal GDP.
Artificial intelligence (AI) has become a very hot topic with incredible recent advances in AI performance. It’s very promising technology, and the expectations shown in thechart above illustrate what would be a profound economic impact. Like many new technologies, however, many find it threatening and are reacting with great alarm, There’s a movement within the tech industry itself, partly motivated by competitive self-interest, calling for a “pause”, or a six-month moratorium on certain development activities. Politicians in Washington are beginning to clamor for legislation that would subject AI to regulation. However, neither a voluntary pause nor regulatory action are likely to be successful. In fact, either would likely do more harm than good.
Leaps and Bounds
The pace of advance in AI has been breathtaking. From ChatGPT 3.5 to ChatGPT 4, in a matter of just a few months, the tool went from relatively poor performance on tests like professional and graduate entrance exams (e.g., bar exams, LSAT, GRE) to very high scores. Using these tools can be a rather startling experience, as I learned for myself recently when I allowed one to write the first draft of a post. (Despite my initial surprise, my experience with ChatGPT 3.5 was somewhat underwhelming after careful review, but I’ve seen more impressive results with ChatGPT 4). They seem to know so much and produce it almost instantly, though it’s true they sometimes “hallucinate”, reflect bias, or invent sources, so thorough review is a must.
Nevertheless, AIs can write essays and computer code, solve complex problems, create or interpret images, sounds and music, simulate speech, diagnose illnesses, render investment advice, and many other things. They can create subroutines to help themselves solve problems. And they can replicate!
As a gauge of the effectiveness of models like ChatGPT, consider that today AI is helping promote “over-employment”. That is, there are a number of ambitious individuals who, working from home, are holding down several different jobs with the help of AI models. In fact, some of these folks say AIs are doing 80% of their work. They are the best “assistants” one could possibly hire, according to a man who has four different jobs.
Economist Bryan Caplan is an inveterate skeptic of almost all claims that smack of hyperbole, and he’s won a series of bets he’s solicited against others willing to take sides in support of such claims. However, Caplan thinks he’s probably lost his bet on the speed of progress on AI development. Needless to say, it has far exceeded his expectations.
Naturally, the rapid progress has rattled lots of people, including many experts in the AI field. Already, we’re witnessing the emergence of “agency” on the part of AI Learning Language Models (LLMs), or so called “agentic” behavior. Here’s an interesting thread on agentic AI behavior. Certain models are capable of teaching themselves in pursuit of a specified goal, gathering new information and recursively optimizing their performance toward that goal. Continued gains may lead to an AI model having artificial generative intelligence (AGI), a superhuman level of intelligence that would go beyond acting upon an initial set of instructions. Some believe this will occur suddenly, which is often described as the “foom” event.
Team Uh-Oh
Concern about where this will lead runs so deep that a letter was recently signed by thousands of tech industry employees, AI experts, and other interested parties calling for a six-month worldwide pause in AI development activity so that safety protocols can be developed. One prominent researcher in machine intelligence, Eliezer Yudkowsky, goes much further: he believes that avoiding human extinction requires immediate worldwide limits on resources dedicated to AI development. Is this a severely overwrought application of the precautionary principle? That’s a matter I’ll consider at greater length below, but like Caplan, I’m congenitally skeptical of claims of impending doom, whether from the mouth of Yudkowsky, Greta Thunberg, Paul Ehrlich, or Nassim Taleb.
As I mentioned at the top, I suspect competition among AI developers played a role in motivating some of the signatories of the “AI pause” letter, and some of the non-signatories as well. Robin Hanson points out that Sam Altman, the CEO of OpenAI, did not sign the letter. OpenAI (controlled by a nonprofit foundation) owns ChatGPT and is the current leader in rolling out AI tools to the public. ChatGPT 4 can be used with the Microsoft search engine Bing, and Microsoft’s Bill Gates also did not sign the letter. Meanwhile, Google was caught flat-footed by the ChatGPT rollout, and its CEO signed. Elon Musk (who signed) wants to jump in with his own AI development: TruthGPT. Of course, the pause letter stirred up a number of members of Congress, which I suspect was the real intent. It’s reasonable to view the letter as a means of leveling the competitive landscape. Thus, it looks something like a classic rent-seeking maneuver, buttressed by the inevitable calls for regulation of AIs. However, I certainly don’t doubt that a number of signatories did so out of a sincere belief that the risks of AI must be dealt with before further development takes place.
The vast dimensions of the supposed AI “threat” may have some libertarians questioning their unequivocal opposition to public intervention. If so, they might just as well fear the potential that AI already holds for manipulation and control by central authorities in concert with their tech and media industry proxies. But realistically, broad compliance with any precautionary agreement between countries or institutions, should one ever be reached, is pretty unlikely. On that basis, a “scout’s honor” temporary moratorium or set of permanent restrictions might be comparable to something like the Paris Climate Accord. China and a few other nations are unlikely to honor the agreement, and we really won’t know whether they’re going along with it except for any traceable artifacts their models might leave in their wake. So we’ll have to hope that safeguards can be identified and implemented broadly.
Likewise, efforts to regulate by individual nations are likely to fail, and for similar reasons. One cannot count on other powers to enforce the same kinds of rules, or any rules at all. Putting our faith in that kind of cooperation with countries who are otherwise hostile is a prescription for ceding them an advantage in AI development and deployment. Regulation of the evolution of AI will likely fail. As Robert Louis Stevenson once wrote, “Thus paternal laws are made, thus they are evaded”. And if it “succeeds, it will leave us with a technology that will fall short of its potential to benefit consumers and society at large. That, unfortunately, is usually the nature of state intrusion into a process of innovation, especially when devised by a cadre of politicians with little expertise in the area.
Again, according to experts like Yudkowsky, AGI would pose serious risks. He thinks the AI Pause letter falls far short of what’s needed. For this reason, there’s been much discussion of somehow achieving an alignment between the interests of humanity and the objectives of AIs. Here is a good discussion by Seth Herd on the LessWrong blog about the difficulties of alignment issues.
Some experts feel that alignment is an impossibility, and that there are ways to “live and thrive” with unalignment (and see here). Alignment might also be achieved through incentives for AIs. Those are all hopeful opinions. Others insist that these models still have a long way to go before they become a serious threat. More on that below. Of course, the models do have their shortcomings, and current models get easily off-track into indeterminacy when attempting to optimize toward an objective.
But there’s an obvious question that hasn’t been answered in full: what exactly are all these risks?As Tyler Cowen has said, it appears that no one has comprehensively catalogued the risks or specified precise mechanisms through which those risks would present. In fact, AGI is such a conundrum that it might be impossible to know precisely what threats we’ll face. But even now, with deployment of AIs still in its infancy, it’s easy to see a few transition problems on the horizon.
White Collar Wipeout
Job losses seem like a rather mundane outcome relative to extinction. Those losses might come quickly, particularly among white collar workers like programmers, attorneys, accountants, and a variety of administrative staffers. According to a survey of 1,000 businesses conducted in February:
“Forty-eight percent of companies have replaced workers with ChatGPT since it became available in November of last year. … When asked if ChatGPT will lead to any workers being laid off by the end of 2023, 33% of business leaders say ‘definitely,’ while 26% say ‘probably.’ … Within 5 years, 63% of business leaders say ChatGPT will ‘definitely’ (32%) or ‘probably’ (31%) lead to workers being laid off.”
A rapid rate of adoption could well lead to widespread unemployment and even social upheaval. For perspective, that implies a much more rapid rate of technological diffusion than we’ve ever witnessed, so this outcome is viewed with skepticism in some quarters. But in fact, the early adoption phase of AI models is proceeding rather quickly. You can use ChatGPT 4 easily enough on the Bing platform right now!
Contrary to the doomsayers, AI will not just enhance human productivity. Like all new technologies, it will lead to opportunities for human actors that are as yet unforeseen. AI is likely to identify better ways for humans to do many things, or do wonderful things that are now unimagined. At a minimum, however, the transition will be disruptive for a large number of workers, and it will take some time for new opportunities and roles for humans to come to fruition.
Robin Hanson has a unique proposal for meeting the kind of challenge faced by white collar workers vulnerable to displacement by AI, or for blue collar workers who are vulnerable to displacement by robots (the deployment of which has been hastened by minimum wage and living wage activism). This treatment of Hanson’s idea will be inadequate, but he suggests a kind of insurance or contract sold to both workers and investors by owners of assets likely to be insensitive to AI risks. The underlying assets are paid out to workers if automation causes some defined aggregate level of job loss. Otherwise, the assets are paid out to investors taking the other side of the bet. Workers could buy these contracts themselves, or employers could do so on their workers’ behalf. The prices of the contracts would be determined by a market assessment of the probability of the defined job loss “event”. Governmental units could buy the assets for their citizens, for that matter. The “worker contracts” would be cheap if the probability of the job-loss event is low. Sounds far-fetched, but perhaps the idea is itself an entrepreneurial opportunity for creative players in the financial industry.
The threat of job losses to AI has also given new energy to advocates of widespread adoption of universal basic income payments by government. Hanson’s solution is far preferable to government dependence, but perhaps the state could serve as an enabler or conduit through which workers could acquire AI and non-AI capital.
Human Capital
Current incarnations of AI are not just a threat to employment. One might add the prospect that heavy reliance on AI could undermine the future education and critical thinking skills of the general population. Essentially allowing machines to do all the thinking, research, and planning won’t inure to the cognitive strength of the human race, especially over several generations. Already people suffer from an inability to perform what were once considered basic life skills, to say nothing of tasks that were fundamental to survival in the not too distant past. In other words, AI could exaggerate a process of “dumbing down” the populace, a rather undesirable prospect.
Fraud and Privacy
AI is responsible for still more disruptions already taking place, in particular violations of privacy, security, and trust. For example, a company called Clearview AI has scraped 30 billion photos from social media and used them to create what its CEO proudly calls a “perpetual police lineup”, which it has provided for the convenience of law enforcement and security agencies.
AI is also a threat to encryption in securing data and systems. Conceivably, AI could be of value in perpetrating identity theft and other kinds of fraud, but it can also be of value in preventing them. AI is also a potential source of misleading information. It is often biased, reflecting specific portions of the on-line terrain upon which it is trained, including skewed model weights applied to information reflecting particular points of view. Furthermore, misinformation can be spread by AIs via “synthetic media” and the propagation of “fake news”. These are fairly clear and present threats of social, economic, and political manipulation. They are all foreseeable dangers posed by AI in the hands of bad actors, and I would include certain nudge-happy and politically-motivated players in that last category.
The Sky-Already-Fell Crowd
Certain ethicists with extensive experience in AI have condemned the signatories of the “Pause Letter” for a focus on “longtermism”, or risks as yet hypothetical, rather than the dangers and wrongs attributable to AIs that are already extant:TechCrunch quotes a rebuke penned by some of these dissenting ethicists to supporters of the “Pause Letter”:
“‘Those hypothetical risks are the focus of a dangerous ideology called longtermism that ignores the actual harms resulting from the deployment of AI systems today,’ they wrote, citing worker exploitation, data theft, synthetic media that props up existing power structures and the further concentration of those power structures in fewer hands.”
So these ethicists bemoan AI’s presumed contribution to the strength and concentration of “existing power structures”. In that, I detect just a whiff of distaste for private initiative and private rewards, or perhaps against the sovereign power of states to allow a laissez faire approach to AI development (or to actively sponsor it). I have trouble taking this “rebuke” too seriously, but it will be fruitless in any case. Some form of cooperation between AI developers on safety protocols might be well advised, but competing interests also serve as a check on bad actors, and it could bring us better solutions as other dilemmas posed by AI reveal themselves.
ImaginingAI Catastrophes
What are the more consequential (and completely hypothetical) risks feared by the “pausers” and “stoppers”. Some might have to do with the possibility of widespread social upheaval and ultimately mayhem caused by some of the “mundane” risks described above. But the most noteworthy warnings are existential: the end of the human race! How might this occur when AGI is something confined to computers? Just how does the supposed destructive power of AGIs get “outside the box”? It must do so either by tricking us into doing something stupid, hacking into dangerous systems (including AI weapons systems or other robotics), and/or through the direction and assistance of bad human actors. Perhaps all three!
The first question is this: why would an AGI do anything so destructive? No matter how much we might like to anthropomorphize an “intelligent” machine, it would still be a machine. It really wouldn’t like or dislike humanity. What it would do, however, is act on its objectives. It would seek to optimize a series of objective functions toward achieving a goal or a set of goals it is given. Hence the role for bad actors. Let’s face it, there are suicidal people who might like nothing more than to take the whole world with them.
Otherwise, if humanity happens to be an obstruction to solving an AGI’s objective, then we’d have a very big problem. Humanity could be an aid to solving an AGI’s optimization problem in ways that are dangerous. As Yudkowsky says, we might represent mere “atoms it could use somewhere else.” And if an autonomous AGI were capable of setting it’s own objectives, without alignment, the danger would be greatly magnified. An example might be the goal of reducing carbon emissions to pre-industrial levels. How aggressively would an AGI act in pursuit of that goal? Would killing most humans contribute to the achievement of that goal?
Here’s one that might seem far-fetched, but the imagination runs wild: some individuals might be so taken with the power of vastly intelligent AGI as to make it an object of worship. Such an “AGI God” might be able to convert a sufficient number of human disciples to perpetrate deadly mischief on its behalf. Metaphorically speaking, the disciples might be persuaded to deliver poison kool-aid worldwide before gulping it down themselves in a Jim Jones style mass suicide. Or perhaps the devoted will survive to live in a new world mono-theocracy. Of course, these human disciples would be able to assist the “AGI God” in any number of destructive ways. And when brain-wave translation comes to fruition, they better watch out. Only the truly devoted will survive.
An AGI would be able to create the illusion of emergency, such as a nuclear launch by an adversary nation. In fact, two or many adversary nations might each be fooled into taking actions that would assure mutual destruction and a nuclear winter. If safeguards such as human intermediaries were required to authorize strikes, it might still be possible for an AGI to fool those humans. And there is no guarantee that all parties to such a manufactured conflict could be counted upon to have adequate safeguards, even if some did.
Yudkowsky offers at least one fairly concrete example of existential AGI risk:
“A sufficiently intelligent AI won’t stay confined to computers for long. In today’s world you can email DNA strings to laboratories that will produce proteins on demand, allowing an AI initially confined to the internet to build artificial life forms or bootstrap straight to postbiological molecular manufacturing.”
There are many types of physical infrastructure or systems that an AGI could conceivably compromise, especially with the aid of machinery like robots or drones to which it could pass instructions. Safeguards at nuclear power plants could be disabled before steps to trigger melt down. Water systems, rivers, and bodies of water could be poisoned. The same is true of food sources, or even the air we breathe. In any case, complete social disarray might lead to a situation in which food supply chains become completely dysfunctional. So, a super-intelligence could probably devise plenty of “imaginative” ways to rid the earth of human beings.
Back To Earth
Is all this concern overblown? Many think so. Bryan Caplan now has a $500 bet with Eliezer Yudkowsky that AI will not exterminate the human race by 2030. He’s already paid Yudkowsky, who will pay him $1,000 if we survive. Robin Hanson says “Most AI Fear Is Future Fear”, and I’m inclined to agree with that assessment. In a way, I’m inclined to view the AI doomsters as highly sophisticated, change-fearing Luddites, but Luddites nevertheless.
Ben Hayum is very concerned about the dangers of AI, butwriting at LessWrong, he recognizes some real technical barriers that must be overcome for recursive optimization to be successful. He also notes that the big AI developers are all highly focused on safety. Nevertheless, he says it might not take long before independent users are able to bootstrap their own plug-ins or modules on top of AI models to successfully optimize without running off the rails. Depending on the specified goals, he thinks that will be a scary development.
James Pethokoukis raises a point that hasn’t had enough recognition: successful innovations are usually dependent on other enablers, such as appropriate infrastructure and process adaptations. What this means is that AI, while making spectacular progress thus far, won’t have a tremendous impact on productivity for at least several years, nor will it pose a truly existential threat. The lag in the response of productivity growth would also limit the destructive potential of AGI in the near term, since installation of the “social plant” that a destructive AGI would require will take time. This also buys time for attempting to solve the AI alignment problem.
In another Robin Hanson piece, he expresses the view that the large institutions developing AI have a reputational Al stake and are liable for damages their AI’s might cause. He notes that they are monitoring and testing AIs in great detail, so he thinks the dangers are overblown.:
“So, the most likely AI scenario looks like lawful capitalism…. Many organizations supply many AIs and they are pushed by law and competition to get their AIs to behave in civil, lawful ways that give customers more of what they want compared to alternatives.”
In the longer term, the chief focus of the AI doomsters, Hanson is truly an AI optimist. He thinks AGIs will be “designed and evolved to think and act roughly like humans, in order to fit smoothly into our many roughly-human-shaped social roles.” Furthermore, he notes that AI owners will have strong incentives to monitor and “delimit” AI behavior that runs contrary to its intended purpose. Thus, a form of alignment is achieved by virtue of economic and legal incentives. In fact, Hanson believes the “foom” scenario is implausible because:
“… it stacks up too many unlikely assumptions in terms of our prior experiences with related systems. Very lumpy tech advances, techs that broadly improve abilities, and powerful techs that are long kept secret within one project are each quite rare. Making techs that meet all three criteria even more rare. In addition, it isn’t at all obvious that capable AIs naturally turn into agents, or that their values typically change radically as they grow. Finally, it seems quite unlikely that owners who heavily test and monitor their very profitable but powerful AIs would not even notice such radical changes.”
As smart as AGIs would be, Hanson asserts that the problem of AGI coordination with other AIs, robots, and systems would present insurmountable obstacles to a bloody “AI revolution”. This is broadly similar to Pethokoukis’ theme. Other AIs or AGIs are likely to have competing goals and “interests”. Conflicting objectives and competition of this kind will do much to keep AGIs honest and foil malign AGI behavior.
The kill switch is a favorite response of those who think AGI fears are exaggerated. Just shut down an AI if its behavior is at all aberrant, or if a user attempts to pair an AI model with instructions or code that might lead to a radical alteration in an AI’s level of agency. Kill switches would indeed be effective at heading off disaster if monitoring and control is incorruptible. This is the sort of idea that begs for a general solution, and one hopes that any advance of that nature will be shared broadly.
One final point about AI agency is whether autonomous AGIs might ever be treated as independent factors of production. Could they be imbued with self-ownership?Tyler Cowen asks whether an AGI created by a “parent” AGI could legitimately be considered an independent entity in law, economics, and society. And how should income “earned” by such an AGI be treated for tax purposes. I suspect it will be some time before AIs, including AIs in a lineage, are treated separately from their “controlling” human or corporate entities. Nevertheless, as Cowen says, the design of incentives and tax treatment of AI’s might hold some promise for achieving a form of alignment.
Letting It Roll
There’s plenty of time for solutions to the AGI threat to be worked out. As I write this, the consensus forecast for the advent of real AGI on the Metaculus online prediction platform is July 27, 2031. Granted, that’s more than a year sooner than it was 11 days ago, but it still allows plenty of time for advances in controlling and bounding agentic AI behavior. In the meantime, AI is presenting opportunities to enhance well being through areas like medicine, nutrition, farming practices, industrial practices, and productivity enhancement across a range of processes. Let’s not forego these opportunities. AI technology is far too promising to hamstring with a pause, moratoria, or ill-devised regulations. It’s also simply impossible to stop development work on a global scale.
Nevertheless, AI issues are complex for all private and public institutions. Without doubt, it will change our world. This AI Policy Guide from Mercatus is a helpful effort to lay out issues at a high-level.
In advanced civilizations the period loosely called Alexandrian is usually associated with flexible morals, perfunctory religion, populist standards and cosmopolitan tastes, feminism, exotic cults, and the rapid turnover of high and low fads---in short, a falling away (which is all that decadence means) from the strictness of traditional rules, embodied in character and inforced from within. -- Jacques Barzun
Sacred Cow Chips 