The marginal propensity to import and multiplier pessimism

This post was written by

In his comment on Kevin’s vertical disintegration post, John Fitzgearld drew attention to the strong linkage between Irish exports and imports.   This got me thinking about Ireland’s marginal propensity to import and its association with, for the want of a better term, multiplier pessimism.  On the face of it, there are certainly grounds for pessimism: in recent years imports have been averaging about 80% of GNP!   But this tells us little about how much of an increase in domestic demand would leak out in increased imports. 

This led me to run an admittedly simple-minded regression of imports on domestic demand and exports.  I use quarterly national accounts data (CSO) for the period 1997:Q1 to 2008Q4 that are in constant 2006 euro and seasonally adjusted.   I run the regression in first differences.   The coefficient on domestic demand is 0.23 (t-stat = 1.9) and the coefficient on exports is 0.60 (t-stat = 5.2); the adjusted R-Squared is 0.48.  [FYI: the same regression in levels yields coefficients of 0.16 (t-stat = 1.9) and 0.68 (t-stat = 10.0)].  

I present these regression coefficients just as suggestive associations in the data.  I’m sure the brainpower on this blog can think of various reasons to doubt a causal interpretation.  However, I find the results plausible given the large share of non-traded services in the economy and the input-output relationships noted by John.   Are we overdoing the multiplier pessimism? 

I hasten to add that I do believe that multipliers are relatively small for an SOE such as Ireland’s; but possibly not that much smaller than the 0.5 that is the rough estimate used by the IMF for SOEs in general.   (See the recent useful primer on multipliers recently released by the IMF’s Research Department.) 

Of course, there are other reasons for multiplier pessimism in the current environment, notably the impact of increased deficits on interest rate spreads and expectations-related effects on current spending.  On the other hand, there are some reasons for (dare-I-say) multiplier optimism: the likelihood that many households and businesses are credit constrained and the fact that monetary policy is almost perfectly accommodative for an SOE that is part of a large monetary union. 

4 Responses to “The marginal propensity to import and multiplier pessimism”

  1. John Fitzgerald Says:

    In ESRI working paper no. 287, which we released in April, we set out some results from using our macro-economic model of the Irish economy – HERMES. In particular, we looked at how a rise in world output would impact on the Irish economy (the results would be similar for a fall in world output of the same magnitude).
    This showed that for a 1% increase in world output, in the long run output in Ireland would rise by betweeen 1,1% and 1.2%. This would be achieved by an expansion in productive capacity in manufacturing and market services resulting in increased exports.
    In this case the rise in the volume of exports would ultimately be around 6.5% and the rise in the volume of imports would be around 5.9%. Almost half of the rise in exports would be exports of services. This increase in the volume of total exports and imports would translate into a rise in imports of 0.73 units for a 1 unit increase in exports.
    This ratio of 0.73 is not the equivalent to a simple multiplier as it takes into account a range of other factors not normally considered in a simple multiplier. In particular, a rise in domestic output, and hence in employment, would tighten the local labour market, causing wages and prices to rise. In turn this would affect competitiveness and, inter alia, the volume of imports.
    This is not the same as a measure of the import content of exports as it includes imports needed to satisfy increased consumption and investment that would result from the shock to world output and hence Irish output.
    In this model simulation of the effects of a rise in world output the increase in the volume of value added in Irish manufacturing (GDP arising) would be around 27% of the increase in gross output. However, not all of the material and services imports would be imported. In addition to the rise in manufactured exports, the rise in services exports would be of a similar magnitude.

  2. Brian Woods Says:

    Just a question.

    China may (or may not!) increase its economic activity. Since they have a large underused labour supply (at all levels – graduate and non-graduate), and a ‘world-class’ manufacturing infrastructure, and a cost-base significantly less than ‘western’ economies, do they really need us at all? That is, to supply them with goods and services, rather being a consumer of their products?

    Comment: If we reduce our cost base to the same level as China, then we can ‘compete’, otherwise not. So what is it that we can export that will be in demand – at our prices? Food? Water? I have real nagging concerns about the next few years, especially if the models being used to suggest possible future economic situations are based on Business-as-Usual (aka. Permagrowth). Is it possible to run a simulation using a GDP value of -0.5% pa? Consumer spending in Ireland is contracting and property asset values are decreasing; what are the negative contributions of these to current GDP?

    Brian P

  3. John McHale Says:

    John, thanks for the reference to the ESRI paper. It is an impressive piece of analysis and should be widely read. Although the regresson I report above is very crude by comparison, the association between exports and imports does appear to be in the same ballpark. If you happen to check back on this thread, I would be interested to see if you think the implied marginal propensity to import out of domestic demand (0.23) is also in the right ballpark.

  4. Geckko Says:

    It would be nice if the great and good of this site gave proper attribution to those hoi polloi who raise issue. You will find this issue was raised by me in an earlier post (with reference to Ireland as a “re-exporting”, rather than a “small open” economy).

    The hypothesis is a sensible one given what we know aboutbthe structure of the economy. A large amount of FDI (relative to domestic output) is parachuted into Ireland and these foreign companies create a flow of production, with very much lower levels of value added than in non-FDI companies. We know this from simple observation of output to labour ratios which show these FDI capital in Ireland to be ludicrously productive.

    If you are well informed, you will also know that much of the operations of the likes of Intel, MIcrosoft, Dell etc. are little more than boxing and shipping.

    The data to test this can be found ni supply and use tables (conveniently 2005 data was recently released) and you find the evidence.

    Go straight to the Use tables and identify the largest exporting industries.

    Then go to the Leontief Inverse tables and find the direct and indrect multipliers for industry groups.

    What you find that the top 10 industry groups account for more than 80% of total exports and have an average direct and indirect import multiplier of 0.58.

    All the remaining industries show a direct and indirect import multiplier of 0.32.

    It is a statistical fact that the exporting industries of Ireland are the industries with the (very high) propensity to import (directly or indirectly).

    As I stated, Ireland is clearly a re-exporting economy, like Belgium, or Hong Kong.

    It is final export demand that leads to high import leakage, not domestic demand.

    I am surprised Brian Lenihan hasn’t had the benefit of anyone in his department pointing that out to him.

Leave a Reply