The regime tile on the Macroeconomic Risk Scorecard currently reads SLOWDOWN, described underneath as the business cycle phase. That is a useful label and it is one dimension. The framework most allocators actually think in has two, because growth and inflation move independently and the combination is what determines which of your holdings is under pressure.
Growth decelerating with inflation falling and growth decelerating with inflation rising are both slowdowns. They are close to opposite environments for a portfolio. One of them is the environment bonds are built for. The other is the one where nothing in a conventional mix works, which is the whole reason the grid exists as a mental model.
Two axes, four cells, and where each axis comes from
The grid is simple enough to draw on a napkin. One axis is the direction of growth, accelerating or decelerating. The other is the direction of inflation, accelerating or decelerating. Direction, not level. That distinction does most of the work, because markets price the change rather than the absolute reading, and a five percent inflation rate that is falling is a different market from a three percent rate that is rising.
Both inputs are things the scorecard tracks. The module monitors CPI and inflation data on one side, and Treasury yields, labour and output series and policy settings on the other, across more than fifty macroeconomic indicators sourced from FRED, the BLS, the BEA and the ECB. What it presents you with on the regime tab is the collapsed version, one phase label. Splitting it back into two axes is something you do yourself, and it takes about five minutes with the inflation data on one tab and the growth-sensitive series on another.
One practical note. You are looking for direction over roughly three to six months, not month to month. A single inflation print that surprises does not flip an axis. Two or three consecutive months moving the same way does.
What each cell does to a stock, bond and crypto mix
The reasoning below is about mechanism, which is to say about why each asset is under pressure in each cell, rather than about what returns happened historically. I will come back to why that distinction matters.
| Cell | What is under pressure and why |
|---|---|
| Growth up, inflation down | The easiest cell for a conventional mix. Earnings expectations rise while the discount rate is not being pushed higher, so equities carry the load and bonds are unremarkable rather than painful. |
| Growth up, inflation up | Equities can hold up on the earnings side while bonds struggle, because rising inflation pushes yields up and existing bonds down. Real assets tend to be where the relative relief is. |
| Growth down, inflation down | The cell bonds exist for. Falling yields lift existing bond prices while equity earnings expectations fall. This is the classic case for a mix that holds duration rather than only cash. |
| Growth down, inflation up | The uncomfortable one. Earnings expectations fall and the discount rate rises at the same time, so equities are squeezed from both ends and bonds do not cushion. Cash stops being a drag and starts being a position. |
Crypto sits awkwardly on this grid and it is worth being direct about that. It behaves mostly like a high-beta risk asset, which means it takes the growth axis hard and is most exposed in the bottom two cells. The story that it is an inflation hedge has been told a lot and the price behaviour has not consistently supported it. Treat it as the most growth-sensitive thing you own and you will be wrong less often than if you treat it as a hedge.
Why I am not giving you a return number for each cell
You will find tables online that tell you exactly what stocks, bonds and crypto returned in each quadrant. I am not going to produce one, and the reason is not caution for its own sake.
First, the sample. Each cell contains a handful of episodes over the whole modern data era, and those episodes differ enormously in policy setting, starting valuation and duration. An average across four or five very different periods is a number with a huge standard error dressed up as a fact.
Second, crypto specifically has existed for a fraction of that history and has lived through very few complete macro cycles. Any per-cell return figure for crypto is computed on a sample so small that it tells you about two or three specific episodes rather than about the cell.
Third, and most practically, the cells are assigned with hindsight. Knowing which cell you were in is easy afterwards and genuinely hard at the time, because the data revises and the direction is only clear once you have enough months to see it. A table of historical cell returns implicitly assumes you knew which cell you were in on day one. You did not.
What survives all three objections is the mechanism column above. Why a rising discount rate hurts long-duration assets is not a statistical claim, it is arithmetic, and it holds regardless of sample size.
Placing today, and the one move it justifies
Today the phase label reads SLOWDOWN, which places you on the decelerating half of the growth axis. That narrows you to the bottom two cells and leaves the inflation direction as the question you have to answer yourself, from the inflation data rather than from the phase label.
If inflation is decelerating alongside growth, the bottom left cell, the argument for holding some duration rather than sitting entirely in cash gets stronger. If inflation is accelerating while growth decelerates, the bottom right, the argument is for cash over bonds and for reducing the most growth-sensitive thing you hold, which for most retail accounts is either a concentrated equity theme or crypto.
Now the important part, which is the size of the move. This is a tilt, not a switch. On a twenty five thousand dollar account, a reasonable expression is moving about five percentage points of the account, twelve hundred and fifty dollars, from the most growth-sensitive holding into cash or short bonds. Not thirty percent. Not everything.
The reason for that restraint is that your placement on the grid is a judgement made from noisy, revising data about a direction that only becomes clear in retrospect, and the combined score sitting at 29 with a LOW risk label is telling you the model set is not alarmed. A modest tilt is what that evidence supports. If the grid placement is right you have improved your position slightly and you can add to the tilt next quarter when it is clearer. If it is wrong, you have given up a small amount of upside and learned something about your own reading, which is a cheap price for a live test of a framework you intend to keep using.