Macro & Strategy - August 2026

Read the full article (PDF)open_in_new

When the Levee Breaks

When excess borrowing meets excess savings

“If it keeps on rainin’, levee’s goin’ to break.”

– Led Zeppelin

For more than two decades, global financial markets have operated behind an enormous levee. Not one made of concrete and steel, of course, but a macroeconomic one: the global savings glut.

The concept is simple but powerful. When desired savings exceed desired investment, interest rates fall until the two sides of the ledger come back into balance. In the early 2000s, then Federal Reserve Governor Ben Bernanke argued that excess savings from emerging Asia, oil exporters and aging advanced economies were flowing into global capital markets, depressing long-term real rates and helping finance large U.S. current account deficits.

Throughout the 2010s, the concept evolved, but the message stayed the same: The world had more savings than investment opportunities, and that surplus kept a lid on interest rates. But the world has embarked on the largest investment cycle in decades. Technology companies are spending unprecedented sums to build artificial intelligence (AI) infrastructure. Governments are ramping up defence spending in response to geopolitical instability. Public investment programs in infrastructure and energy security are expanding.

Each initiative may be justified on its own merits. Together, they represent an extraordinary demand for capital. But observers may rightfully wonder whether markets will be able to absorb all the debt.

The good news is that the levee is still standing. But the pressure on the system is rising.

The identity behind it all

The starting point is the humble macroeconomic identity: Savings equal investment. In practice, the world isn’t a closed spreadsheet, and the adjustment happens through prices, currencies, capital flows and expectations. But, at the global level, the intuition is clear. If there is more desired saving than desired investment, the price of capital falls. If desired investment rises faster than savings, the price of capital rises. Interest rates are the balancing item.

Highlights

• The era of abundant, cheap capital is fading; a powerful global investment cycle is underway as governments and corporations adjust to a new technological and geopolitical landscape.
• The most likely outcome is a regime of structurally higher long-term rates, wider differentiation between winners and losers, and a less forgiving environment for weak borrowers.
• This investment cycle supports an overweight in equities and commodities, while supporting an underweight position in fixed income.

The idea that excess global savings hold interest rates down entered the economic mainstream in the mid-2000s. At that time, the explanation helped solve a puzzle: Why were longterm interest rates so low despite solid economic growth and rising government debt?

The answer was that desired savings had grown faster than desired investment.

Emerging economies accumulated reserves after the crises of the 1990s. China became the world’s manufacturing hub and generated enormous external surpluses. Aging populations saved for retirement. Corporations became increasingly conservative after the global financial crisis. Investment remained subdued as savings continued to rise.

In other words, money became abundant.

When savings consistently exceed investment opportunities, capital becomes cheap. Real interest rates decline. Financial assets become more valuable. Governments discover that deficits can be financed more easily than expected.

This world appeared to have changed in 2022, when the savings glut became a shortage. Inflation forced central banks to tighten aggressively. Real rates moved back into positive territory. Households spent their pandemic-era excess savings. Russia’s invasion of Ukraine accelerated defence spending in Europe and elsewhere. The arrival of generative AI launched a new capital expenditure race. And governments, already carrying heavy debt loads, kept on borrowing.

Still, as the chart above shows, the situation rapidly normalized, and from 2023 onwards, savings continued to exceed investment. As a result, interest rates, although higher than in the 2010s, remain low by historical standards. Term premia, the extra compensation investors demand to hold longer-dated bonds instead of rolling over short-term debt, have risen from their Covid-era lows but are still below most levels seen in recent decades. Credit spreads also remain historically tight, pointing to investors’ continued willingness to absorb a broad range of debt.

In short, the world still operates with excess savings. If capital were to become genuinely scarce, the implications for financial markets would be significant. The next sections examine the key trends that will determine whether today’s global savings glut could turn into a global savings shortage.

The investment boom: AI, defence and the state

The AI arms race

The biggest new source of investment demand is artificial intelligence. The numbers are large enough to matter at the macro level.

AI investment is surging, with estimates constantly rising. The latest projections point to $800 billion in capital expenditures in 2026, with cumulative spending potentially reaching $4 trillion by 2030 and creating a global demand shock for all sorts of electrical components, energy, commodities and, perhaps most importantly, credit.

At first, the AI buildout was mostly self-financed. The tech giants generated enough free cash flow to fund aggressive investment programs while still supporting buybacks. That is one reason credit markets have remained calm.

But the trend is changing. The sector’s investment needs are outstripping internal cash generation. Tech sector financing is shifting toward external markets, with projections for net bond issuance rising meaningfully through 2030.

AI-linked companies are in fact already on a bond binge. At mid-year, hyperscalers had issued $250 billion in bonds, more than double the total for 2025¹. To fund their investments, they’re even going beyond the U.S. market. Alphabet and Amazon recently issued bonds in the Canadian market, successively notching record bond sales by a single corporate issuer in this country. The issuances were well received by the Canadian markets, but the room for further absorption is limited.

The other arms race

The public sector is the second major source of investment demand. Military spending declined globally from 1980 to 2010 and then flatlined for a decade, but is rising again across most economies. The IMF reports that half of the world’s countries increased their defence spending from 2020 to 2024, with more to come. NATO countries have pledged to spend 5% of GDP on defence by 2035, a significant increase from current levels.

Infrastructure spending is also rising as governments try to rebuild supply chains, support industrial policy and adapt to geopolitical fragmentation. For example, in Europe, government investment as a share of GDP is rising steadily, after staying at a low level for more than a decade. Germany in particular is keen to upgrade its aging infrastructure to jump-start a stalled economy.

The story is similar in Canada, where Mark Carney’s government is boosting infrastructure spending to support overseas trade and rebuild the military. Unlike transfer payments, which often move savings from one sector to another, public investment absorbs resources directly. It requires labour, materials, equipment and financing.

Much of this spending will be financed by debt issuance. The IMF projects that global public debt will cross 100% of GDP by 2029, up 6% from 2025 levels.

The supply of bonds is, therefore, rising from both sides: sovereigns and corporations. The OECD estimates that governments and corporations will borrow $29 trillion in 2026, roughly double the amount of 10 years ago. And, as a share of GDP, after flatlining during the 2010s, bond issuance is on the rise again after the Covid pandemic.

That’s a lot of paper for the market to digest. The good news is that there is plenty of savings to absorb this increasing issuance.

The savings side: still large but less unconditional

The savings glut hasn’t disappeared. Worldwide, gross domestic savings were about 27% of GDP in 2024, slightly below 2022 levels, and the trends have been mostly stable in recent years in the major economies.

Europe and Asia still save a lot. China still runs a large external surplus, as do Japan, South Korea and Taiwan. Wealth inequality continues to concentrate savings among households with high saving propensities. Corporate savings remain healthy outside the technology sector. In a more uncertain world, precautionary savings may even rise in some regions.

But one pillar of the old savings structure is eroding: emerging market reserve accumulation. After a string of crises in the 1990s, central banks in emerging markets started building large reserves, recycling their external surpluses into U.S. Treasuries and other safe assets. That is changing.

Central bank foreign exchange reserves have been shrinking as a share of world GDP since 2015. Reserve managers net sold $210 billion of bonds in the first half of 2025. They have shifted their reserve composition toward gold, reducing the price-insensitive support that once helped anchor global bond markets.

That change is crucial. A central bank that buys Treasuries for reserve management isn’t the same buyer as a hedge fund, an exchange-traded fund investor or a household saver. The first buyer is often less price-sensitive. The second buyer wants compensation. The market can still clear, but the clearing yield is higher.

In brief, the world isn’t running out of savings. Rather, the world is running out of the old kind of savings: large, patient, official, price-insensitive and automatically recycled into safe bonds.

Can markets absorb the debt?

We have little evidence that global credit markets are facing an immediate absorption crisis. Corporate credit spreads are still contained. Investment-grade markets continue to finance AI-related issuers. Sovereign auctions have become more sensitive but not dysfunctional. Equity markets have also shown capacity to absorb large issuance, especially when the story is tied to structural growth.

But, with the growing rate of investments, along with the changing composition of savers, the global savings glut may not be the shock absorber that it used to be. Three broad outcomes are possible.

In the first and most likely scenario, global savings continue to grow but at a slower pace than investments. Markets absorb the additional supply, but it steadily pushes real interest rates higher. In this scenario, long-term yields remain elevated. The world funds the investment boom, but the cost of capital becomes more prohibitive, forcing investors and policymakers to become more selective about what they choose to finance.

A second possibility is that investment demand proves even stronger than expected. Artificial intelligence spending could continue, surprising to the upside. Defence spending could expand further, while fiscal consolidation remains politically difficult.

In that environment, the rain intensifies and the pressure on the levee increases. Higher real rates and higher term premia would be required to restore balance but would cause turmoil in financial markets. Such a scenario could even trigger a debt crisis for some heavily indebted governments.

A third scenario would see the AI investment boom fizzle out. In that world, the savings glut reasserts itself and long-term yields resume their downward trend.

Investment implications

For investors, the message is subtle but all-important. It isn’t a call to abandon bonds. Higher yields mean bonds once again provide income. That’s good news for balanced portfolios and income-oriented investors. But it’s difficult to build a strong case for a sustained decline in long rates unless the investment boom breaks or global growth deteriorates materially.

The second implication is dispersion. When savings are abundant and central banks are buying, almost everyone gets funded. When savings are still available but less plentiful, investors discriminate. Sovereigns with weaker fiscal trajectories pay more. A case in point is the United Kingdom: Its sovereign bond yields are higher than those of any other G7 country, for fiscal unsustainability and political instability are a toxic mix for investors.

Corporations with credible cash flows and strategic assets can still borrow cheaply. More speculative borrowers may face a higher hurdle. Credit spreads are still low, but a widening is likely over the medium term.

Conclusion: The future can be funded but at a cost

The world isn’t facing an imminent funding crisis. The pools of savings are still vast, and global capital markets remain deep. The AI buildout, defence spending and the infrastructure cycle can continue.

But the era of effortless absorption is fading. The old savings glut was supported by reserve accumulation, corporate cash hoarding, weak investment and central bank bond buying. Today, reserve managers are less supportive, governments are dissaving, the technology sector is spending aggressively and private investors are the marginal buyers of duration. They’re willing to buy, but they aren’t doing charity work

That’s why long rates should stay high and may even go higher. Central banks can cut policy rates, but they can’t manufacture a 2010s-style savings glut if investment demand is rising and fiscal supply remains heavy.

It’s a world where capital has a price again. Sound projects will be funded. Weak credits will pay up. Governments will face more scrutiny. Investors will earn income, but duration will need a clearer growth scare to deliver strong capital gains.

In other words, the levee hasn’t broken, but it’s getting more expensive to maintain.

Positioning

We remain overweight equities, with continued optimism for Canadian equities. We believe Canada is well positioned to benefit from the combination of accelerating global capital expenditure, resilient nominal growth, and strengthening demand for commodities and real assets, while maintaining attractive relative valuations versus several major developed markets. Despite these trends, market participants remain pessimistic on Canada due to trade tensions with the United States. More broadly, our constructive stance on equities reflects strong corporate profitability and a powerful investment cycle centered on artificial intelligence infrastructure, data centers, electrification, and strategic industrial investment.

We remain underweight fixed income. Although bond yields are substantially higher than they were during the previous decade, the environment remains challenging for duration. Fiscal deficits remain elevated across much of the developed world, sovereign borrowing requirements continue to grow, and competition for capital is increasing as both governments and corporations seek financing for large-scale investment programs. In addition, inflation remains above target in many economies, limiting the ability of central banks to aggressively ease policy. While tactical opportunities may emerge, we continue to view the strategic risk-reward profile of government bonds as unattractive relative to other asset classes.

We maintain an overweight allocation to commodities. The structural backdrop remains highly supportive as economies and corporations continue to compete for access to energy, industrial metals, and other strategic resources. The ongoing buildout of physical infrastructure associated with artificial intelligence, electrification, defense spending, and supplychain resiliency is creating durable demand for real assets. At the same time, geopolitical fragmentation and a greater emphasis on economic self-sufficiency are encouraging inventory accumulation and investment in domestic production capacity, both of which support commodity demand. Commodities also continue to serve an important role as inflation-sensitive assets in a world characterized by expansionary fiscal policy, elevated debt burdens, and recurring supply disruptions.

1. Includes Alphabet, Meta, Nvidia, Oracle and SpaceX. Source: Wall StreetJournal.