Mike Alexander - The K Cycle
The Kondratiev Cycle and Secular Market Trends
In my introduction to secular market trends I mentioned in passing that beginnings of secular bull markets (troughs in the stock cycle) were closely related in timing to the Kondratiev Cycle. In my recent "Irrational Exuberance" article I mentioned that the timing of the stock cycle was determined by the Kondratiev cycle. In this article I introduce the Kondratiev cycle to those investors unfamiliar with it, and show how the secular market trends are related to it. This article is simply intended to be an overview. The reader is referred to my book Stock Cycles for a more detailed treatment of this subject matter.
The Kondratiev cycle is a 50-60 year cycle in prices, interest rates and other economic variables. It was noted as early as 1847 in an article in the British Railway Journal by Dr. Hyde Clark, but it was N. D. Kondratiev who first described the cycle in detail and for whom the cycle is named. Kondratiev cycles are most readily apparent in monetary data such as prices and interest rates. Figure 1 shows a plot of the US producer price index over the period 1800 to 2000. Prior to WW II, the Kondratiev cycle could be seen clearly as periodic peaks in prices spaced about 50 years apart.
Figure 1. The U.S. producer price index 1800-2000
Figure 1 shows three "Kondratiev peaks" in producer prices in 1814, 1864, and 1920. In between the peaks are two "Kondratiev troughs" in prices, one in 1843 and the other in 1896. The spacing of these peaks and troughs average 53± 3 years, which provides an estimate for the length of the Kondratiev cycle. This observation, in a nutshell, provides the original empirical basis for the Kondratiev cycle. We call the period of rising prices, between the Kondratiev trough and the Kondratiev peak, the Kondratiev upwave, or just upwave. Conversely, we call the decline from the peak to the trough the downwave.
If we examine each downwave closely we see that after each Kondratiev peak there is a sharp drop and then a leveling-off in prices, producing what is sometimes called the (price) plateau. The plateau ends with a second precipitous drop, or what is sometimes called the fall from plateau. This second drop bottoms in what has been called the vortex by the economist Brian Berry. The first three cycles following 1800 showed plateaus ending in 1818, 1873, and 1929. The vortices following the fall from plateau occurred in 1830, 1878 and 1932. These points are marked in Figure 1.
Following the vortex, there is a temporary rise in prices to an intermediary peak, before prices fall still further to the Kondratiev trough. Following Berry, I call the price peak reached at this time the "deflationary growth peak" or DG-peak. DG-peaks occurred in 1836 and 1882. These peaks in prices were close to the ends of downwave secular bull markets that had begun around the time of the Kondratiev peak. Thus, associated with the first Kondratiev cycle, there was a Kondratiev peak in 1814, and a downwave secular bull market from 1815 to 1835, which ended close to the DG peak in 1836. Associated with the second Kondratiev cycle, there was a Kondratiev peak in 1864, and a downwave secular bull market from 1861-1881, which ended close to the DG peak in 1882. Associated with the third Kondratiev cycle, there was a Kondratiev peak in 1920, and a secular bull market from 1921-1929. In contrast with the first two cycles, the 1920's bull market ended with the fall from plateau. There was no obvious DG peak and the whole pattern of Kondratiev cycles in prices broke down.
Kondratiev, working in the 1920's, predicted a deflationary depression using his cycle theory. He was right, but subsequent attempts to apply his theory have been unsuccessful due to the absence of deflation (falling prices) after 1932. The disappearance of deflation over the past 70 years is the result of stimulatory economic policy that began with the New Deal and continues to this day. Stimulatory policy falls into to two categories, fiscal and monetary. Fiscal policy involves government expenditures, while monetary policy affects the money supply. By running a deficit the government can put more money into the hands of consumers than it takes in taxes, which can increase the amount of money available for spending. The Federal Reserve can increase the amount of money in circulation by a variety of means. Both of these policies are stimulatory in that they can increase the rate of economic growth (at least in the short term) and produce rising prices.
Figure 2. Prices, U.S. government debt + money supply, and reduced prices for 1800-2000
Figure 2 shows the trend in stimulation in terms of the sum of government debt and money supply as a percentage of real GDP . Along with this measure of stimulation (bold black) is plotted the price index (blue), both on a logarithmic scale. Note that the shape of both lines is similar and that the trend towards higher prices mirrors the trend towards higher debt + money. An empirical model for prices was constructed using linear regression . This model gives what the price "should be" if fiscal/monetary stimulation were the only factor that affected prices. The actual price is usually different from the predicted value. The ratio of the actual price to the predicted price is plotted in Figure 2 as the reduced price (red line). The reduced price tracks the changes in price level after the complicating effect of fiscal/monetary stimulus has been removed. We see the same pattern of Kondratiev peaks and troughs in the reduced price over the entire 200 year period as we saw in the raw price index before 1932.
What this means is the same Kondratiev cycle is operative today as was operative in the 1920's, when Kondratiev first described it in detail. We see peaks in reduced price in 1813, 1864 and 1918 (Figure 2), that closely correspond to the Kondratiev peaks in raw prices in 1814, 1864 and 1920 (Figure 1). We see two troughs in reduced price after each peak that correspond to the vortex and Kondratiev trough in the raw prices. In between each pair of troughs is a peak (labeled in green) that corresponds to the DG peak in the raw price plot. Note that 1946 is clearly identified as a Kondratiev trough in reduced prices, as it is the second trough after the 1918 peak and is followed by a long-term rise in reduced prices. In the raw price plot, the period around 1946 was a time of high inflation, there certainly was no trough in raw prices at this time.
We also note that 1981 is identified as a Kondratiev peak in the reduced price plot. In contrast, the raw price plot shows an uninterrupted rise throughout the 1970's to 1990's. The peak in 1981 means that since 1981 the economy has been in a downwave, despite still rising prices. Figure 2 shows that the economy has received strong stimulus throughout this downwave, first with the deficit spending of the Reagan and Bush administrations and then by the growth in M3 during the 1990's. As a result of all this stimulus, deflation has not occurred, yet the existence of the downwave is revealed by the reduced price. We see a trough in reduced price in 1992 that might correspond to the vortex for this downwave. If this is the case, the "plateau" for this downwave was very short, even shorter than that for the first cycle. Alternately, we may still be on a very long "plateau" and a further sharp drop in reduced price to a future vortex is coming.
Based on the last three cycles we should expect a DG peak to occur some 17-22 years after the Kondratiev peak, or in the 1997-2002 time period. The first two DG peaks in 1836 and 1881 occurred close to the ends of downwave secular bull markets in 1835 and 1881. The identification of 2000 as the end of the most recent downwave secular bull market is completely consistent with a DG peak around now. This interpretation supports the idea that the vortex is already past. Alternately, the 2000 stock peak may presage the fall from a very long plateau, just as the 1929 peak was associated with a fall from plateau.
Either interpretation suggests that reduced prices should peak soon, very possibly this year. What it does not say is that actual inflation will peak. Although our analysis of the Kondratiev cycle shows that we have a decade or more of downwave left, all this tells us is that reduced price will fall from the levels today to a second trough. What could easily happen is stimulus could sharply increase, like it did in the 1940's. This would result in falling reduced prices even though actual inflation might increase from today's level. The very rapid reduction in the Federal Funds rate and President Bush's much touted tax cut could combine to send stimulation sharply upward in the coming years, producing strong "deflation" in the reduced price measure without the same in actual prices.
The analysis I presented shows 1981 as the most recent Kondratiev peak, a finding consistent with that of David Knox Barker his 1995 book The K-wave. Other interpretations are possible, (see Shilling) most common is the view that the most recent Kondratiev peak was in 1974 rather than 1981, in which case the Kondratiev trough should be close approaching or perhaps even already here (see Drake).
So far we have discussed the Kondratiev peak solely in terms of prices and derived measures. Figure 3 shows a comparison of reduced prices with interest rates on high grade corporate bonds, return on resources (ROR) and capital accumulation. Recall that ROR is simply S&P500 earnings divided by business resources (E/R) and capital accumulation is the annual growth rate in R. The reader is referred to my previous article for a discussion of these quantities and what they mean.
Figure 3. Comparison of reduced prices, with interest rates, ROR and capital accumulation
Figure 3 shows that interest rates show Kondratiev peaks and troughs at around the same times as do reduced prices. The twenty year trailing average of ROR and capital accumulation also show Kondratiev peaks and troughs. Since the fundamentals that drive stock prices (earnings and interest rates) show Kondratiev cycles, it is perfectly natural that that the stock market do so too. Each Kondratiev peak is associated with the beginning of a downwave secular bull market. Hence the 1814 Kondratiev peak is associated with the start of the 1815-35 secular bull market. The 1864 Kondratiev peak is associated with the 1861-81 secular bull market. The 1920 Kondratiev peak was immediately followed by the 1921-29 bull and the 1982-2000 secular bull market followed just one year after the 1981 Kondratiev peak. Each Kondratiev trough is associated with the start of an upwave secular bull market. Thus upwave secular bull markets began in 1843, 1896, and 1949, right around the time of Kondratiev troughs in 1842-43, 1895-1902, and 1946-50. We can expect the next secular bull market to be an upwave bull and to get underway around the time of the next Kondratiev trough.
Each upwave and downwave has its own secular bear market. Thus, two pairs of secular bull and bear market trends (two stock cycles) fall into one Kondratiev cycle. This correspondence provides a stock-based definition of the Kondratiev cycle as a pair of adjacent stock cycles. Thus, we can track the length of the Kondratiev cycle from the 1802-1853 pair of stock cycles to the 1929-2000 pair as shown in Figure 4. There is some evidence of cycle lengthening, but this finding is heavily dependent on the 1929 secular bull market peak that terminated an unusually short secular bull market. Recall that, unlike the previous two downwave secular bull markets (which lasted 20 years and ended at the DG peak) the 1921-29 secular bull market ended at the fall from plateau and lasted only 8 years. The 18 year length of the most recent secular bull market is more consistent with first two downwave secular bull markets than with the third one. Similarly, the interpretation that the vortex was in 1992 and we are at the DG peak now is also consistent with the first two Kondratiev downwaves. If this interpretation is correct, we should expect reduced prices to gradually work lower to a Kondratiev trough bottom in about a decade, and a new secular bull market to begin a few years after that.
Figure 4. Length of the K-cycle (as defined by the stock cycle) over time
On the other hand, if the cycle lengthening is real, this would imply that downwave secular bull markets now end at the end of the plateau (like 1929) and that we should expect a large drop in reduced prices in the near future. If this is correct, the vortex is still ahead and the Kondratiev trough (and beginning of the next secular bull market) might be as far away as 2020. If we do indeed fall into recession later this year or next (as seems likely) we shall see whether reduced prices fall dramatically or not, and so obtain a bearing on our current position within the cycle.
14 May 2001
Mike Alexander, author of
Stock Cycles: Why stocks won't beat money markets over the next 20 years.
1] Government debt was calculated by summing government deficits over time. For money supply, M2 was used between 1869 and 1959. M3 was substituted for M2 in 1959 and afterward (M2 and M3 were about the same in 1959). M2 values are unavailable before 1869. From 1834 to 1869, the sum of currency and bank deposits was used to estimate M2. Before 1834, total currency was used.
2] One hundred years of price data (blue line in Fig 2) were regressed against the corresponding 100 years debt + money data (black line in Fig 2). The regression equation was used to calculate what the price level "should" be for the year in the middle of the regression sample. That is, the model price for 1850 used the regression equation obtained from the data between 1800 and 1900. The regression model used to calculate a value for 1860, used data from 1810 to 1910, and so on. Model values after 1950 used the regression equation obtained from the 1900-2000 data, and values before 1841 used the equation obtained from the 1791-1891 data.
© Copyright 2002 Robert Paterson.
Last update: 30/12/2002; 9:17:23 AM.