Market CommentsBefore I get to the inflation analysis, I want to comment on the market action behind the charts above. To be perfectly clear, I am making comments based on media chatter, and extrapolating the experience of 2008. I have no background sources giving me juicy information.
The first thing to note that the indexed yield (typically called the "real yield", but I have pedantic objections to using "real yield" for the quoted yield on inflation-linked bonds) has been jumping around far more than the breakeven inflation rate. This is exactly what we should expect: the breakeven inflation rate is the true valuation metric, and the real yield is the residual. (This is discussed in Section 4.2 of my book, Breakeven Inflation Analysis (link to Books2Read page). If any of my readers are stuck in self-isolation, I would remind you that my books are attractively priced, and available as e-books.
This is not to say that breakeven inflations will exactly match forecast inflation right now. This is exactly what happened in 2008 (as I discussed in Section 3.5). Inflation-linked bonds are de facto risk assets that can be purchased by many government bond funds. Overweighting linkers is a classic way for bond investors who are bearish on rates to position without blowing up other risk limits, and one of the adorable things about bond fund managers is that the vast majority of them are permanently bearish on rates. Meanwhile, levered long positions in inflation-linked bonds was a core strategy for "risk-parity" portfolios.
Another comment I have seen was to the effect that "breakevens have nothing to do with inflation, it is just oil." This is a puzzling sentiment, given the high sensitivity of headline inflation to gasoline prices in the United States. (I discuss this in Section 3.2, "The Role of Energy Prices".) I think the idea is that people want the market to price core inflation, but that would require a different investment product.
Short-Term Inflation Forecasting
The first thing I noted is that I am skeptical about our ability to forecast the overall economy on an unconditional basis using any mathematical model. (Yes, that includes stock-flow consistent (SFC) models, which are the subject of another book that I am plugging.) The best we can hope for is to do conditional forecasts: given some scenario, what will happen to the economy? The current situation is a very good example: we could only have forecast the upcoming collapse in activity if we had a very good idea about the spread of COVID-19. Although a mathematical epidemic model might give insights, that model is not a traditional economic model.*
However, I carved out an exception for "short-term" CPI forecasting. We can quite often build a pretty good forecast for CPI inflation, conditional on gasoline prices. The various components of CPI are relatively slow-moving, and respond with a lag to underlying factors. We can stitch together partial models, and get a good estimate of the overall CPI. Obviously, these partial models break down over longer forecast horizons.
Why I'm Not Worried About InflationAlthough I am wildly uncertain about outcomes, I do not see CPI inflation as being a particular issue right now. The main reason is that in order to get an inflationary spiral, we need wages to rise. Although there are some pockets where workers are getting wage boosts (as well they should, as they are the front line workers taking the greatest risk), this will not happen across the board with expected double digit unemployment rates.
Anyone who looks at central bank balance sheets ("the money supply") is simply going to get wildly wrong inflation forecasts. These are just shifts in holdings of financial assets, and mean diddly-squat about the real economy. The only significance they have is relative to the counter-factual scenario where the central bank lets the financial system implode in a debt death spiral. Inflation is higher relative to that counter-factual, but it does not mean that prices are going to go up.
As for the fiscal response, some will just be used to fill in cash flow gaps of firms (and line rich people's pockets). This does not put demand pressure on anything (other than relative to the total implosion counter-factual). The rest is income replacement households, and I see no appetite for this to be a permanent universal basic income in any jurisdiction.
A temporary income replacement is just preventing the deflationary liquidation of the household sector. Once again, this is only "inflationary" relative to a disastrous counter-factual. (Right now, I would argue that the Canadian and American responses are too small, but that might be rectified as reality bites politicians in the nether regions.)
Even if money is sent to households that do not really need it, they cannot spend it on much of anything. Everything except grocery stores and pharmacies are closed, and physical delivery infrastructure is going to be maxed out. There is no sign of price gouging by physical retailers. Everybody with a couple brain cells to rub together knows that this is a shared event, and will be seared into everyone's memories for a very long time. Running around acting like a sociopath -- a course taken by a few idiot employers -- is going to be remembered by a lot of people after the dust settles.
As I will discuss later, some prices will be going up, but they are unlikely to be enough to move the CPI by a lot.
The easiest thing to buy are electronic downloads (like my books!). Since the marginal cost of production is nil, the usual strategy is to keep pricing fixed, and hope for a volume increase.
Supply Chain Horror StoriesThe big potential scare story is the collapse in supply chains, and rising import costs. I have some sympathies for that story, but it is somewhat self-limiting. Right now, on-shoring production is catching a lot of interest. Overseas suppliers starting a price-gouging campaign would put populist protectionism movements on steroids.
CPI WeightingI will now run through the sectors of the United States CPI. I will not attempt to give forecasts, but give a qualitative outline of my thinking. Although there are a few areas where price behaviour are different, it would be a similar story in at least some other developed countries (Canada and the United Kingdom at least).
I am using the weights from the February 2020 U.S. CPI report: https://www.bls.gov/news.release/archives/cpi_03112020.htm. I will keep the categories in the original order, which may make the discussion somewhat jumpy, and round off to one decimal place. I have skipped some smaller categories.
- Food (13.8%). This one of the few categories where we will actually have legitimate retail prices, given the shutdown of other sectors. However, note that "Food" is not just groceries: it is split between food at home (7.6%), and away from home (6.2%). Restaurants will be struggling to survive, and so it is hard to imagine them ramming through price increases. As for food at groceries, a significant portion is processed, and thus prices are administered. Supply bottlenecks will almost certainly appear in fruit and vegetables that rely on masses of workers for harvest, but that categories is only 1.3% of the CPI. Grains require low levels of labour input, and energy costs are extremely important (fuel and fertiliser).
- Energy (6.7%). Retail prices will not fall as much as wholesale crude prices, but the collapse in crude prices will blow a big hole in CPI. Outside the U.S., the larger share of gasoline taxes will reduce the deflationary shock in the CPI.
- Household Furnishings and Supplies (3.7%). People stuck at home might see more things to spend on, but hardly a high priority given income uncertainty most households face.
- Apparel (2.8%). People under stay-at-home orders are probably not worried about fashion trends. Office workers working from home just need a good shirt to give a veneer of presentability.
- New Vehicles (3.7%): Automakers are desperate to survive. Nobody can go anywhere.
- Used Vehicles (2.5%): Hawking a second car is probably going to be a standard way to raise cash.
- Medical Care Commodities (1.6%): (Note: "commodities" = goods in CPI jargon.) There are some obvious shortages, but once again, most firms do not want to be seen as high profile sociopaths. The exception are pharmaceutical firms, but their prices are normally hidden behind the veil of insurance.
- Recreation Commodities (2.0%). Probably some demand here, at least for delivered items.
- Education Commodities (0.5%). Probably increased interest, but prices are administered.
- Alcoholic beverages (1.0%). Low energy prices will reduce price pressure on grain inputs and processing, and bars and sporting venues will be shut for a long time, blowing a hole in demand.
- Tobacco and smoking products (0.6%). Let's smoke in the midst of a killer pandemic that goes after weak lungs in a grisly fashion!
- Shelter (33.1%). A category with a huge weight and extremely slow-moving price dynamics. Split between rent (7.8%) and owner's equivalent rent (24.0%). Rent contracts are negotiated on an annual basis. and we have a massive income shock to renters. Very hard to see how rents go up. In any event, will be extremely slow-moving, so even if rents somehow went up, won't show up for awhile.
- Medical Care Services (7.2%). I leave researching what will happen here to the reader. I think this is largely mediated by insurance, and I would not hazard a guess as to what the numbers will be.
- Transportation Services (5.4%). Travel is not exactly a hot spot for price pressures.
- Recreation Services (3.7%). Largely digital, and follow the "keep prices stable, go for volume" strategy.
- Education and Communications (6.4%). What happens with education is anybody's guess, but it is hard to see how hefty fees can be charged for remote education services. Communication services are a huge growth area right now, but once again, they are likely to aim for increased volume at unchanged prices.
- Other personal services (1.7%). Largely the domain of small businesses, that are under severe pressure.
In summary, although we can see some problem areas, without wage pressures hitting across the board, hard to sound an alarm about inflation.
* There are various heterodox academics that insist that economics ought to be multi-disciplinary, and include things like epidemic models. Although I can agree with the sentiment, it is very difficult to see how such models can be calibrated to data. Why not include asteroid strike scenarios? Killer bees? Etc.
(c) Brian Romanchuk 2020