Economic cycles: Kitchin, Juglar, Kuznets, the Kondratiev wave and the investment clock

An economic cycle is a recurring swing in output, investment and prices around a long-run trend. Over the past 160 years economists have described four cycles of different lengths, each tied to a different driver: business inventories (about 3–4 years), equipment investment (7–11 years), construction and housing (15–25 years), and waves of new technology together with long swings in interest rates and debt (50–60 years). The investment clock, a framework published by Merrill Lynch in 2004, links the shorter cycles to growth and inflation. None of them is a timetable. They are ways of reading the background in which every trade is made.

The Kitchin cycle: inventories (3–4 years)

In 1923 the British statistician Joseph Kitchin found that US and UK prices, interest rates and bank clearings rose and fell in cycles of about 40 months.

• Driver. Companies order stock based on expected sales. When demand surprises on the upside, shelves empty, orders and production rise, and inventories are rebuilt; when demand slows, unsold stock piles up and orders are cut until it is worked off. The swing in production is larger than the swing in final demand, which is why it shows up in growth figures.
• Four phases. Analysts often split it into passive destocking (sales recover faster than stock), active restocking (companies rebuild on purpose), passive restocking (sales slow and stock accumulates) and active destocking (companies cut orders).
• Data to watch. The Census Bureau's total business inventories/sales ratio, published monthly in the Manufacturing and Trade Inventories and Sales report, and inventory changes in the GDP accounts. A falling ratio with rising sales points to the early phases; a rising ratio with slowing sales points to the later ones.
• Limits. Just-in-time production and a larger service sector have made inventory swings smaller than in Kitchin's day, and supply shocks (2021–22) can move the ratio for reasons that have nothing to do with demand. The reported values are in current dollars, so inflation flatters sales growth.

The Juglar cycle: business investment (7–11 years)

The French physician and economist Clément Juglar described in 1862 a recurring sequence of prosperity, crisis and liquidation in France, Britain and the United States, lasting roughly 7 to 11 years.

• Driver. Companies invest in machinery and equipment when demand and profits are strong and credit is cheap. New capacity takes years to build and pay off; when it overshoots demand, or credit tightens, investment falls sharply. Most economists now treat this as close to the ordinary business cycle that the National Bureau of Economic Research dates through its expansions and recessions.
• Data to watch. Real nonresidential fixed investment in the Bureau of Economic Analysis GDP release, new orders for nondefense capital goods excluding aircraft (the Census Bureau's durable goods report), capacity utilization (Federal Reserve), and lending standards in the Federal Reserve's Senior Loan Officer Opinion Survey.
• Limits. Lengths vary widely between cycles, and policy (interest rates, fiscal support) can lengthen or cut short an expansion. Recent capital spending is concentrated in data centers and software, which the older indicators measure less well.

The Kuznets cycle: construction and housing (15–25 years)

Simon Kuznets, later a Nobel laureate, identified in 1930 swings of 15 to 25 years in production and prices, which later work linked to building activity, migration and population growth.

• Driver. Housing and infrastructure take years to plan and build and last for decades. Waves of household formation and immigration create demand; construction responds with a lag, often overshoots, and is then followed by a long slow period. Because housing is financed with long-term debt, mortgage rates amplify the swing.
• Data to watch. Housing starts and building permits (the Census Bureau and the Department of Housing and Urban Development, monthly), new and existing home sales, and mortgage rates.
• Limits. The evidence is mixed. In 1965 the economist Irma Adelman found little support for regular long swings in US data, and statisticians have shown that the moving averages used to reveal such cycles can create apparent waves in random data (the Slutsky–Yule effect). Housing cycles are real, but their length is far from regular.

The Kondratiev wave: technology, interest rates and debt (50–60 years)

In the 1920s the Russian economist Nikolai Kondratiev argued that prices, interest rates and output in the industrial economies moved in long waves of roughly 50 to 60 years. Joseph Schumpeter later tied each wave to a cluster of innovations and named the cycles after Kitchin, Juglar and Kondratiev; Carlota Perez extended the idea to technological revolutions and the financial bubbles that accompany them.

• Driver. A new general-purpose technology spreads through the economy: investment and productivity rise, and so eventually do interest rates and prices. As the technology matures, returns fall, debt built up in the boom weighs on growth, and rates decline until the next technology starts a new upswing. Each wave is usually split into recovery, prosperity, recession and depression.
• Commonly cited waves (dates differ between authors): steam power and textiles from the late 18th century; railways and steel from the mid-19th century; electricity and chemicals from the end of the 19th century; the automobile and petrochemicals from the mid-20th century; information technology from the 1970s and 1980s.
• Data to watch. Long-term Treasury yields over decades, public and private debt relative to GDP, productivity growth, and the spread of a new technology through investment.
• Limits and controversy. This is the most disputed of the four. Two centuries of data hold only four or five waves, too few to test statistically; authors disagree on the turning points by a decade or more; and much of the original evidence came from prices, which central banks now manage. Many mainstream economists regard the long wave as a useful story rather than a regular cycle.

Where it might stand now. A widely cited view, popular in China, holds that the fifth (information technology) wave peaked in the boom before the 2008 financial crisis, and that the world has since been in its recession and depression phases, with a new upswing waiting on technologies such as artificial intelligence and new energy. Others date the turning points differently or doubt that they can be dated at all. Two measurable facts sit behind the debate: the 10-year Treasury yield was 4.96% on October 22, 2026 (US Treasury daily par yield curve), well above the lows of 2020, and total US public debt outstanding was $40.41 trillion on October 21, 2026 (US Treasury, Debt to the Penny). Whether higher rates mark the start of a new long upswing or the strain of a late one is exactly what the theory cannot settle. Our reading: neutral, because the long wave cannot be timed.

The investment clock

Merrill Lynch's "The Investment Clock" (2004) sorts the economy by two questions: is growth above or below its trend, and is inflation rising or falling? Using US data from 1973 to 2004, it found that different asset classes had tended to do best in each of the four resulting phases:

• Data to watch. Real GDP growth (BEA, quarterly) and the jobs report for growth; CPI and core CPI (Bureau of Labor Statistics, monthly) for inflation; the federal funds rate (Federal Reserve) for policy.
• Limits. The phases are only clear in hindsight, the clock can skip phases or run backwards, and the asset ranking is an average over one period of history, not a rule. Since 2004 the relationship between stocks and bonds has also changed more than once.

The latest official data (as of October 22, 2026). Real GDP grew at an annual rate of 2.2% in the second quarter of 2026 (BEA, third estimate, September 30, 2026). Payrolls rose by 29,000 in September 2026 (preliminary) and the unemployment rate was 4.2% (Bureau of Labor Statistics). CPI was up 3.7% from a year earlier in September 2026, up from 2.4% in January and February, while core CPI, which leaves out food and energy, was up 2.6% (Bureau of Labor Statistics, October 2026); see the CPI page. The rise in headline inflation has come mainly from items outside the core, while core inflation has held steady. The Federal Reserve raised the federal funds target range by 0.25 percentage point to 3.75%–4.00% on September 17, 2026; see the FOMC page.

How Clear Tape reads the clock. Clear Tape's big-cycle indicators apply fixed rules to official monthly data. Growth is the average six-month annualized growth of the four coincident indicators that the National Bureau of Economic Research watches when it dates recessions (payrolls, industrial production, real personal income excluding transfers, and real manufacturing and trade sales), compared with its median over the past ten years. Inflation is the change in the core PCE price index over the last three months, annualized, compared with its change over the last twelve months. For July 2026, the latest month with all the data, growth was 1.94% against a trend of 1.69%, 0.25 percentage point above it, and core PCE rose at a 2.30% annualized rate over three months, 0.68 point below its 2.98% twelve-month rate. Growth above trend with core inflation cooling places the US in recovery, with medium confidence; a new phase is shown only after it has held for two months in a row. The core measure is what separates this reading from a glance at headline CPI: headline inflation has risen mainly on items outside the core, while the momentum of core inflation is slowing. Our reading: neutral to optimistic, with the risk that the rise in non-core prices spreads into core inflation.

On the shorter cycles, the same indicators read the inventory cycle from two directions: the year-over-year growth of manufacturers' new orders and of business inventories, each as a three-month average and compared with three months earlier. Both were rising in August 2026, which the rules call active restocking, with medium confidence. The Census Bureau's business inventories/sales ratio fits: 1.33 at the end of August 2026, down from 1.40 a year earlier, while sales were up 10.2% in current dollars (Census Bureau, October 15, 2026). Reading: neutral to optimistic, allowing for inflation in the sales figures. Housing starts ran at a seasonally adjusted annual rate of 1,221,000 in September 2026, 7.4% below a year earlier, with permits 3.0% lower (Census Bureau and HUD, October 20, 2026). Reading: cautious.

The member center shows the latest cycle position, updated monthly: Big-cycle indicators.

What it means for a trader

• Background, not a signal. Cycles describe the conditions that most stocks trade in: rates, earnings growth and which sectors lead. They are far too slow and too imprecise to time an entry or exit.
• Test rules across different conditions. A rule that has only been tested in one phase may stop working when the phase changes. Testing it over long, different periods, for example a run of years with two recessions and a later run with a pandemic and an inflation shock, shows whether it depends on one kind of market. Backtesting and overfitting explain how.
• Adjust risk, not conviction. When the background turns less favorable, the practical lever is smaller position sizes and fewer open trades, not a prediction of where the market goes next.
• Watch the releases that move the cycle. CPI, the jobs report and FOMC decisions reset rate expectations quickly; the economic calendar lists their dates, and how FOMC and CPI move stocks covers the trading side. Market breadth shows whether a move is broad or narrow.

In Clear Tape

The market data pages show the latest CPI, jobs and FOMC figures from the official sources, free. The big-cycle indicators in the member center show where the investment clock and the inventory cycle stand, by the rules above. In the trading-system backtest, every rule is judged on 2000–2014 and then run unchanged on 2015 to now, two periods that cover very different parts of the cycle. Writing the macro background into each plan in the trade journal, such as "inflation rising, rates up", lets you see after a few months whether your results depend on it.

The readings in this article (optimistic, neutral, cautious) describe the official data at the date given. They are not recommendations to buy or sell anything.

For educational purposes only. Not investment advice.