Double Dip? ECRI at the Crossroads

Economic Cycle Research Institute’s leading economic indicators have been a matter of spirited debate the last couple of weeks, because it looks like they are on the verge of calling a double dip recession. Yet no word from ECRI, as the pressure increases for them to make a call. The interest in their stance on the economy is compounded by their correct call for an end to this past recession at the depth of the crisis last year.

First let’s clear up some misinformation about ECRIs leading index, known as WLI.

They do not just rely on one index called the Weekly Leading Index, WLI. It is actually four indexes of which WLI is the most visible. The others are WLI Growth, the Long Leading Index, USLLI, and USLLI Growth. The growth indexes are derived from their counterparts. ECRI uses a combination of these indexes to make their calls, so it is not technically correct to take the WLI and make a call from the one index no matter how sharp its decline.

However if the index were to move decisively by itself, it could act as an indicator without the aid of its sister indexes. There is no reason for ECRI to use it that way though since they get more information from all the indexes combined, facilitating earlier and more accurate calls. Since ECRI is the only one that knows all the moving parts it is a difficult to disagree with them. The problem for the rest of us is we may wish to make our own calls when it seems obvious to us that a call should be made.

Currently WLI Growth is at -9.8% having fallen many weeks in a row. There has always been a recession when WLI Growth falls below -7.0. Certainly there is a number low enough for WLI to signal recession without the other supporting indexes and without ECRI making their own call. Is it -10 or -15 or some other number? Just how much extra cushion is needed is not clear.

The appeal of using just WLI and WLI Growth is clear. They are the only indexes ECRI regularly releases to the public. The growth index is a little more of a mystery since its calculation is not disclosed. There is almost no information available on USLLI and USLLI Growth.

Interestingly, WLI, the most widely known and best understood appears to be the least useful in the analysis. The least known, USLLI Growth appears to be the most useful, at least to the outsider. ECRI has a valuable franchise to protect so it makes sense for them to protect the key ingredients.

The only reasons to make your own projection are if you believe ECRI has a reason to withhold or suspend making a decision, or you have good reason to not want to wait and are willing to assume a greater risk your own forecast goes awry. Certainly ECRI can have legitimate business reasons to hold back on the forecasts as long as possible, to avoid mistakes.

For an investor once the odds of a correct forecast go over 50% there is every reason to want to get the forecast as soon as possible, especially if the stock market has not adequately discounted the forecast yet. ECRI on the other hand may want to wait until the odds of a correct forecast are much higher to maintain or nurture their reputation.

In the current situation WLI Growth is at -9.8% and there has always been a recession with the index at that level. The index is currently trending down sharply and consistently. WLI has also trended downward at a sharp rate of decline. The problem for ECRI in making a forecast is USLLI has only recently turned downward. ECRI management wants to hold off until the trend is more pronounced.

At this point if one does not wish to accept ECRIs timetable it is time to do your own research into ECRI. Trying to build your own set of similar indexes is one way, but a very challenging project. Another more practical way is to use their numbers and develop your own protocols for them, against the set of recession data that is available, that is for recessions from 1970 forward.

In building your own model the goal is to find one set of principles that apply, accurately forecasting all the recessions in a meaningful timeframe. In starting that process and comparing the numbers for WLI, WLIG and USLLI against the start of the seven recessions since 1970 there are three recessions that stand out as possible problems.

Two of them, the ’91 and ’01 have the similar problem of a very short lead time for LLI to make the recession call. Since ’01 has both a shorter lead time and the trend is much shallower we will concentrate on it as the most problematic in building a model and making a correct recession call. The ’82 recession has a different problem raising questions about ECRIs protocols not directly related to the current situation.

The problem with incorporating the ’01 recession into a model is USLLI’s peak was only 5 months before the decline began. That leaves a maximum of 5 months of data on LLI to establish the downtrend necessary to make the call. Based on an evaluation of the data, the natural time to make the call would have been on 12/29/00 with just a little over two months of data and a fairly weak downtrend. USLLI dropped only about a point and a half or about 1% over that two month time span. That may not be pronounced or persistent enough to make the call. Yet to delay further would have meant to watch the data deteriorate calling into question the basis for the call. The next time the data improves enough to make a stronger call it is already late March after the 2001 recession has already started. Remember in constructing a model we want to make these calls with the benefit of hindsight with the understanding they have to stand up under objective, not subjective interpretation.

The problem for the here and now is this. ECRI has not made a call yet on what may be a coming recession, based on a lack of confirmation from USLLI. As it turns out it now has 2.5 months of downtrend since the peak of USLLI in the Spring. That’s more time than when ECRIs data made the call in ’01. So, either ECRI believes the call should have been made closer to the beginning of the recession in 2001 when the data temporarily turned against the call, or after the recession started in March, or on 12/29, with two months of data. It has to be one way or the other Either the call was made in a shorter time frame or it was made under conditions adverse to the model.

So the question is why haven’t they made the call yet? Why do they need more time this time than in 2001, especially when WLI and WLI Growth are so clear in their signal? Has the data turned around in the last week? I doubt it.

In order for ECRI to maintain internal consistency, specifically with respect to ’01, they have a limited window in which to make the recession call this time around, or else invite questions about the consistency and objectivity of their model. It is for this reason I expect their call very soon, in the next couple weeks.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: