Because the literature is divided over the issue of whether military spending stimulates economic growth or not, my defining work must explain why that perception exists and in fact what is the real nature of military spending that produces this apparent conflict in the literature.  My own exhaustive studies of the long term empirical record, mostly of the last century in America but also with important international checks, clearly show the following results.

Appearance of Local Economic Growth

Locally and regionally, it is clear that economic activity is added to localities where military spending is spent.  That economic activity is similar to manufacturing and other goods producing activities in the normal economy like construction, mining, and agriculture.  Both kinds of economic activity consume large amounts of capital, engineering, and scientific research. 

Looking Closer

The local economic growth from military spending is also clearing shown in the US historical state by state record to come with a similar depletion of economic growth in civilian manufacturing and lower economic growth in states and regions high in manufacturing and low in military spending.  So the local economic growth from military spending represents a transfer of resources from low military spending states to high military spending states.  This last point shows up in studies of changes in the per capita military spending by state from year to year in the US Statistical Abstract on the allocation of federal spending.  I have been studying these changes many different ways for many years now, and there is a strong pattern in the states of manufacturing going down as military spending goes up across the many state economies.  In fact, in the top 20 states, the correlation of military spending with the government sector of each state is about 0.83.  So I can say that military spending usually represents the major part of government spending changes state by state.

War Claim:  Military Spending Stimulates Growth

The economy often increases in growth in war years.  When military spending goes down, at the end of a war, the economy often goes into recession.  This would give a correlation of growth with military spending.

Looking Closer

Government borrowing increases substantially during wars, and then the debt is repaid when peace comes.  The federal deficit expands enormously during war periods compared to other periods of time. My essentially perfect long term sixty year model of the US economy is predicated on the simple assumption that deficits raise the economic growth rate of each year and military spending decreases the economic growth rate of each year.  Thus, net military spending, that is military spending minus deficit spending, is the amount of slowing force in the economy.  So the war economy phenomenon is better explained as a federal deficit economy.  Comparing manufacturing productivity growth rates with economic growth rates does suggest some additional war boom that can't be explained away.  I call this the adrenaline rush boost of a war time situation, also consistent with the post war let down.

Research Claim:  Military Spending Stimulates Growth

In the US case, research has historically been about 10% of the military budget.  Hence this research case for military spending can at most be 10% true for military spending as a whole.  Take the internet claim as an example.  Some say the race to put a man on the moon laid the groundwork for the internet.  British and Swiss scientists claim credit for important parts of the internet, so it's not just DARPA in the USA that can claim credit for the internet.  Analysts find a much reduced impact of military research by dividing the impact three ways, one for being too military specific, one for secrecy reducing the benefit of interaction with other scientists, and one for dual purpose civilian and military usage of inventions.  Hence 33% effective is a common estimate of the military research impact on the economy.  One pro military buildup expert claims the impact even lower, 7%.  That makes sense in the venture capitalist estimate that the inventor only gets 20% of the benefit as competitors get the other 80% benefit of a new product.  Hence 20% of the 33% estimate leaves the 7% estimate for US benefit. Since that 7% applies to the 10% research part of the budget, then military spending is 99.3% not stimulative.

Roosevelt and War

The economy grew 86% in the eight prewar years from 1933 to 1941 as Keynes advised Roosevelt to spend money and he ran a combined 31% deficit in those years.  That's an average growth rate of almost 11% per year with a mostly jobs program deficit of 4% of the economy per year.  The unemployment rate dropped from 25% in 1933 to 10% in 1941.  Happy days were here again.  Note how the New Deal got triple the growth rate from their deficits.  The war did fully employ people again, with an average unemployment rate of about 2%.  Note that both periods, New Deal and war, reduced unemployment about 2% per year.  But look how inefficient the deficit was with military deficits.  The four war years grew 26% on 155% deficits.  That's 6.5% growth on 39% deficits on average.  So the multiplier under New Deal programs was 2.75 and under the war was 0.17.  That makes the peacetime deficit multiplier 16 times more effective that the war time deficit multiplier.  Giving all the credit to the war for ending the Depression is to mistake an accident of historical timing for cause and effect.  Military spending is an extremely inefficient way to stimulate the economy.

Summary

Two different twenty year international developed world comparisons by Ruth Leger Sivard suggest that military spending represents lost capital investment and lost manufacturing productivity growth rate in the economy, while clearly giving a local economic boost.  But closer inspection always shows that military spending positives are offset in other parts of the economy.  The sizes of military increases in wartime or peacetime are always greatly in excess of the so-called economic growth benefits.  Deficits stimulate economic growth and military spending is often a part of that deficit, but the offsets in national capital investment and manufacturing productivity growth show up clearly in local and national statistics.  Hence military spending is a popular excuse to avoid paying taxes to balance off spending increases.  Yet military spending is a very inefficient way to provide stimulus.  Hence the truism, military spending is non productive.

For 10 key "defining" statistical models of Military Economics:

https://www.academia.edu/4044456/SUMMARY_Military_DisEconomics

For a complete hundred year war and peace economy history of America since 1910:

https://www.academia.edu/4044532/HISTORY_Presidents_Military_Economy_1910-2009

For more on the regional economic nature:

https://www.academia.edu/5740273/MIDWEST_and_the_Military_3_pages_2005