Thursday, January 27, 2022

Data Update 3 for 2022: Inflation and its Ripple Effects!

    Inflation numbers have been coming in high now, for more than a year, but for much of the early part of 2021, bankers, investors and politicians seemed to be either in denial or casually dismissive of its potential for damage. Initially, the high inflation numbers were attributed to the speed with the economy was recovering from COVID, and once that excuse fell flat, it was the supply chain that was held responsible. By the end of 2021, it was clear that this bout of inflation was not as transient a phenomenon as some had made it out to be, and the big question leading in 2022, for investors and markets, is how inflation will play out during the year, and beyond, and the consequences for stocks, bonds and currencies.

Inflation: Measurement and Determinants

    As the inflation debate was heating up in the middle of last year, I wrote a comprehensive post on how inflation is measured, what causes it and how it affects returns on different asset classes. Rather than repeat much of that post, let me summarize my key points.

  • Measuring inflation is not as simple as it looks, and measures of inflation can vary depending on the basket of good/services used, the perspective adopted (consumer, producer, GDP deflator) and the sampling used to collect prices. That said, the three primary inflation indices in the US, the CPI, the PPI and the GDP deflator all told the same story in 2021:
    Download historical inflation numbers

    The inflation rate during the course of the year reached levels not seen in close to 40 years, with every price index registering a surge.
  • While news stories focus on reported (and past inflation, it is expected inflation that should drive investment, and measures of these expectations can come from surveys of consumers (University of Michigan) or from the market, as the difference between the treasury bond rate and the inflation-protected treasury bond, of equivalent maturity:
    Download data

    Using the ten-year bond, it is clear that while inflation expectations have inched up in the bond market, but that rise is far more muted than in the actual inflation indices, and consumer expectations of inflation now significantly exceed the bond-market imputed estimate for expected inflation.
  • While the implied inflation in bond rates is low, investors seem to be anticipating higher inflation. Using a measure that the Federal Reserve has developed, I report the percent of investors expecting inflation to be greater than 2.5%, representing one end of the inflation expectation spectrum, and those expecting deflation, representing the other, in the graph below:
    Download data
    As you can see the 93.96% of investors were expecting inflation to be greater than 2.5%, by the end of December 2021, up from 6.74% in December 2020, suggesting a sea change in the market. Conversely, the percent of investors expecting deflation has dropped to a vanishing low number, suggesting that Cathie Wood has little company, in her contention that deflation is the real danger to markets and economies.
The undeniable fact is that inflation came back in 2021, but the question of why it happened, and whether it will stay high, is  hotly debated. To those who believe that it is a spike that will dissipate over time, it is another casualty of COVID, as a combination of virus-driven supply chain issues and government spending to offset shutdowns has driven prices up. In this mostly benign story, inflation will go back down, once these pressure ease, though it is unclear to what level. To others, and especially those old enough to remember the 1970s, it does seem like a return to more unsettled times, with potentially dangerous consequences for the economy and markets.

Interest Rates and Inflation

    Inflation and interest rates are intertwined, and when their paths deviate, as they sometimes do, there is always a reckoning. While we have increasingly given central banks primacy in discussions of interest rates, it remains my view that markets set rates, and while central banks can nudge market expectations, they cannot alter them. Put simply, no central bank, no matter how powerful, can force market interest rates down, if inflation expectations stay low, or up, if investor are anticipating high inflation. 

US Treasuries: A Mostly Uneventful Year

    After a turbulent year in 2020, when COVID shut the global economy down, and interest rates plunged and stayed down for the rest of the year, 2021 was a more settled year, with long term rates rising gradually over the course of the year, but short terms rates staying put:

Treasury Rates Data

While treasury bills continued to yield rates close to zero, rates increased for longer term treasuries, with 2-10 years rates rising much more than rates on the longest term treasuries (20-year to 30-year). For those who track the slope of the yield curve, and I am not one of those who believes that it has much predictive power, it was a confusing year. The treasury curve became steeper, but only at the shortest end of the spectrum, with the slope rising for the 2-year, relative to the 3-month, but not at all, when comparing the 10-year to the 2-year rate. Beyond the 10-year maturity, the slope of the yield curve actually flattened out, with the difference between the 30-year rate and the 10-year rate declining by 0.34%.

Corporate Bonds: No Shortage of Risk Capital

    In my last post, I chronicled the movement in the equity risk premium, i.e. the price of risk in the equity market, during 2021, but the bond market has its own, and more measurable, price of risk in the form of corporate default spreads. Using bond ratings classes to categorize companies, based upon credit risk, I looked at the movement of default spreads during 2021:

Download data

Corporate default spreads decrease across ratings classes, but the decline is much larger for lower rated bonds, with the default spread on high yield bonds registering a drop of 1.25%. Note that the decrease in default spreads, at least for the lower ratings, mirrors the drop in the implied equity risk premium during the course of 2021. Read together, it suggests that private risk capital continued to not just stay in the game, but increased its stake during the course of the year, extending a decade-long run.

Expected Inflation, Interest Rates and Bond Returns

    While day to day movements in interest rates are driven by multiple forces, including the latest smoke signals coming from central banks and investor sentiment, the longer term and drivers of interest rates are fundamental. In particular, if you start by breaking down a long term riskfree rate (like the 10-year treasury bond) into an expected inflation and an expected real interest rate components, you can also reconstruct an intrinsic risk free rate by assuming that the real growth in the economy is a stand-in for the real interest rate and that most investors form expectations of future inflation by looking at the inflation in the most recent year(s):

Download data

In this picture, the actual ten-year treasury bond rate is superimposed against a rough measure of the intrinsic risk free rate (obtained by adding together the actual inflation rate and real growth rate each year) and a smoothed out version (where I used the average inflation rate and real growth rate over the previous ten years). Not only has the intrinsic risk free rate moved in sync with the ten-year bond rate for most of the last seven decades, but you can also see that the main reason why rates have been low for the last decade is not the Fed, with all of its quantitative easing machinations, but a combination of low growth and low inflation. Coming into 2022, though, the intrinsic risk free rate is clearly running ahed of the ten-year treasury bond rate, and if history is any guide, that gap will close either with a rise in the treasury bond rate or a decline in the risk free rate (coming from a recession or a rapid drop off in inflation). 

Unexpected Inflation and Asset Returns

    Note that it is expected inflation that drives interest rates, and that the actual inflation rate can come in above or below expectations. In my post on inflation last year, I drew a contrast between expected and unexpected inflation, arguing that financial assets are affected differently by each component. If expected inflation is high, but it is predictable, investors and businesses have the opportunity to incorporate that inflation into their decision making, with investors demanding higher interest rates on bond and expected returns on stocks, and businesses raising prices on their products/services to cover expected inflation. Unexpected inflation is what catches us off guard, with unexpectedly high inflation leading to a reassessment of pricing (for all financial assets) and an uneven impact across businesses, leaving those with pricing power in a better position than those without that power. 

    To assess how inflation has affected asset returns over time, I broke down the actual inflation rates since 1954 into expected and unexpected components each year, using a brute force assumption that the average inflation rate over the last ten years is the expected inflation rate. (In the last two decades, we have had access to more sophisticated measures of expected inflation, including the difference between the nominal treasury bond and TIPs rates, but not in earlier years). In the graph below, I look at annual returns on stocks, treasury bonds and corporate bonds, with the unexpected inflation numbers also shown:

There are a few aspects of this graph that stand out. With my crude measure of inflation expectations, it looks like it takes time for inflation expectations to shift, during periods of higher or lower inflation, as can be seen in the extended stretches of higher than expected inflation, in the 1970s, and lower than expected inflation in the 1980s. 

    While it is not immediately visible in the graph, returns on stocks and bonds are affected by unexpected inflation, and to illustrate by how much, I broke the 94 years of data into five quartiles, based upon the level of unexpected inflation, with the lowest (highest) quintile representing the years when inflation came in most below (above) expectations, and estimated the annual returns (nominal and real) for stocks, treasuries and corporate bonds in the table below:

With treasuries and corporates, the returns generally get worse, as inflation comes in above expectations, with real returns showing the damage from unexpected inflation. With equities, the sweet spot in terms of returns is when inflation is at or below expectations, and the worst scenarios are when inflation comes in well above expectations.  I also looked at how inflation plays out on equity sub-groupings, on two dimensions, the first being market capitalization and the second being price to book, with the former becoming a stand-in for the vaunted small cap premium and the latter for the value versus growth question.

Over much of the last century, small cap stocks have done better than large cap stocks, when inflation has come in well above expectations, perhaps providing some insight into why the vaunted small cap premium seems have to disappeared over the last two decades of muted inflation. Similarly, the value effect, computed as the premium (or discounted) return earned by low price to book stocks over high price to book, becomes more pronounced during periods when inflation is greater than expected and much less so, during periods when inflation is lower than anticipated.
    Using the same approach with gold and real estate, with the caveat that historical data on the former is more limited, I get the following results:
While gold and real estate both do better than financial assets, when inflation is greater than expected, there is also a clear difference between the two investment classes. Real estate operates more as a neutral hedge, delivering returns that are, for the most part, unscathed by unexpectedly high inflation, but gold is a bet on inflation, delivering the highest returns, when inflation is much greater than expected, and negative returns, when it is lower than expected. Much as I would like to extend this analysis to newer investment classes, there is not enough historical data on crypto currencies or NFTs to allow for the analysis. As I noted in my inflation post in 2021, though, the early evidence is not promising for these new investment categories, at least as inflation and crisis hedges, since they have behaved more like risky equities, at least on a day-to-day basis and during the 2020 crisis, than like gold.

Inflation and Currencies
    Much of this post has been about inflation in the US, and by extension, in US dollar terms, it is worth emphasizing that inflation is a currency-specific phenomenon. While inflation in the US dollar, by dint of its status as the currency in which commodities are priced, can sometimes spill over into other currencies, it remains true that you can have high inflation in one currency, while there is low inflation in other currencies. Inflation differences across currencies play out in two domains, with the first being interest rates in different currencies and the other being exchange rate.
Interest Rates across Currencies
    I start every one of my discussions of discount rates with a truism, by stating that the riskfree rate that you start with should reflect the currency in which you have decided to do your valuation. That then becomes the springboard for estimating risk free rates in different currencies, following one of two paths. In the first, you start with government bond rates in the local currency, in different currencies, and adjust those rates for default risk in the local currency government bond. (Government bonds in local currencies do default, and account for a significant proportion of sovereign defaults in the last 50 years). My estimates for the start of 2022 for the currencies where local-currency government bonds are available is below:
Download data

Riskfree rates are highest in currencies, like the Zambian Kwacha or Turkish Lira, where inflation is highest, lower in low-inflation currencies and even negative in currencies, where deflation may be the long term prediction. I am using the default spreads based upon the local currency sovereign ratings for the countries in question, with the government bond rate being the risk free rate only for currencies where the issuing government in triple-A rated. If you dislike this assumption, or do not believe that the government bond rate is a market-set number in a particular market, there is a second approach, where you start with the risk free rate in US dollar or Euros, and adjust it for differential inflation, i.e., the difference in expected inflation between the US and the country in question:
Thus, if the US treasury bond rate is 1.5%, and expected inflation rates in the US and Indonesia are 1% and 4% respectively, the approximate riskfree rate in Indonesian Rupiah will be 4.5% (=1.5% + (4%-1%)) and the more precise riskfree rate in Rupiah will be 4.52% (=1.015*(1.04/1.01)-1). While the expected inflation rate in dollars may be an easy get, it is more difficult to get expected inflation rates in other currencies, but the IMF has estimates for the next five years at this link.

Exchange Rates
    Just as interest rates in currencies are determined, in large part, by inflation differentials, exchange rates over time are also driven by those same inflation differentials. Drawing on one of the oldest relationships in exchange rates, purchasing power parity, you can extract the forward exchange rate in a currency:

Thus, currencies with higher inflation can be expected currency devaluation over time, relative to currencies with lower inflation. As with interest rates, in the short term, there are forces, ranging from central banking intervention to momentum and speculation, that can cause rates to deviate from the inflation script, but in the long term, it is almost impossible to break the cycle.
    Connecting this linkage to the discussion of US inflation in the prior sections, here are the takeaways. If you believe that inflation will stay high, not just in the US, but across the globe, the exchange rate effects will be muted. If, on the other hand, you believe that the inflation shock will vary across countries, your actions will be more nuanced. For the countries where you believe that local inflation will decrease, relative to the US,  the US dollar will weaken against their currencies, augmenting returns you will earn in their markets (stock or bond). For countries, where you see local inflation surging more than you expect to see in the US, the US dollar will strengthen against their currencies, reducing the returns you make in their markets. As with the discussion of asset returns, it is not expected inflation that is the source of exchange rate risk, since you can incorporate those expectations into exchange rates, but unexpected inflation, which, when extreme, can cause significant revaluations of currencies. 

    As with any historical data assessment, I could give you the standard boilerplate disclaimer that past performance is not always a good predictor of the future, but to the extent that the past provides signals, your expectations of how inflation will play out in the coming year will play a key role in your asset allocation and stock selection decisions. If you believe that last year's surge in inflation is a precursor to a long time period when inflation is likely to stay high, and come in above expectations, you should be shifting your holdings away from financial to real assets, and within your equity holdings, towards small cap stocks,  stocks trading at lower pricing multiples (PE, Price to Book) and companies with more pricing power. If, on the other hand, you believe that inflation worries are overdone, and that there will be a reversion back to the low inflation that we have seen in the last decade, staying invested in stocks, and especially in larger cap and high growth stocks, even if richly priced, makes sense.

YouTube Video

Data Links
Blog Post Links


Anonymous said...

Hi Professor - in your Riskfree Rate graph, why is the 10-yr bond rate (red line) lower than the intrinsic risk free rate (black line)? I think I inherently don't understand how the intrinsic rate is being calculated. thank you!

Manas Mehrotra said...

Dear Prof.Ashwath,

the second half of your post reminded me of Prof.Jayanth Varma's class on Forward Rates and currency valuation. I could actually visualize myself teleported and sitting in his class as I read aloud your post.I adore both of you and vastly respect your art and knowledge.

I must congratulate and thank you for your clinical approach in slicing out each topic and taking a scientific and nuanced approach instead of a soothsayers one.

I love your posts and look forward to them eagerly. Would you care to write a post on the current IPO boom and the subsequent dampener that took place recently in the Indian stock markets? How do you analyze the same?

Thanks & Regards,

Manas Mehrotra